When most people propose changes in our electoral system, they generally do so in order to achieve a political end, not because the change conforms to a platonic ideal of what elections should be like. So it is with the plaintiffs in Evenwel v. Abbott, No. 14-940, which the Supreme Court will hear this fall. Their claim is that, when states draw their legislative districts, the Equal Protection Clause requires that they use the numbers of voters, instead of the number of people, as the basis for allocating seats within the states.
The Supreme Court has ended the most blatant forms of gerrymandering and required legislative districts at both the state and federal level to be equal in composition within each state. The Court's rulings have been labeled "one person, one-vote," and the general assumption has been that, in dividing up each house by districts, the denominator has been the total population of the state.
Evenwel challenges that assumption and argues that, because the goal of one person, one vote is to have each person's vote count the same as every other person's, the denominator should be total voters and not total population. If this were the law, the main groups that would no longer be counted are children, illegal immigrants, those not registered to vote, and felons who are precluded from voting. Until the actual lines are drawn for all the districts in a state, the results are not certain. But we do know that the backer of this lawsuit (Edward Blum) also supports Fisher v. University of Texas, which seeks the elimination of affirmative action in university admission. Therefore, it is reasonable to assume that he believes that the change would have an adverse impact on minorities and their Democratic supporters, or at least it has that potential in some states, including Texas where the case was brought.
As a constitutional challenge, the case could establish one of three legal rules: the first, urged by the plaintiffs, is that the Constitution requires that the denominator be comprised of voters only; the second, argued by Texas, is that the states are free to choose among several reasonable options for the denominator; and third, urged by no party to the case, but perhaps will be advanced by the Obama Administration, is that the Constitution requires that the base be total population for all state elections. Nothing in the language of the Constitution helps answer that question.
When the Supreme Court stepped in to stop gerrymandering, it did so to impose law on state legislators, instead of allowing them to use their unbridled discretion on how to do re-districting, in large part because they stood to gain (or lose) depending on what changes were made. The state-choice position of Texas is essentially an argument that there is no constitutional law to apply.
However, the Court has made clear that equality is required, subject to certain permitted deviations, which cannot be squared with the claim of Texas that each state can decide what base it will use to allocate its legislative districts. In fact, until 1981, when the Texas Attorney General opined that the provision of the Texas Constitution that apportioned the State Senate based on "qualified electors" was unconstitutional, Texas used that alternative basis. Thus, if Texas "wins" this case on its theory that the choice is up to the States, Texas could change back to a "qualified elector" based system, and plaintiffs would have achieved a victory although losing in the Supreme Court.
By contrast, the plaintiffs in Evenwel argue that the total voters in a state is the right choice, which at least has the benefit of recognizing that the question is a legal one, for which there is a right answer. But there is an important wrinkle to their position. In asking the Court to hear the case, they agreed that, while the challenged districts are equal in population, they have "gross disparities in voters or potential voters" (emphasis added). That final phrase highlights a serious practical problem with plaintiffs' position: how to define "voters" let alone "potential voters."
There is a simple and agreed upon answer for population: the decennial official census, mandated by Article I, section 3 of the Constitution. Because there is no constitutional definition for "voters," it could mean any or all of the following: registered voters; voters who voted in the last election; voters who voted in the last presidential election; voters who voted for state legislators in a recent election; or registered voters plus those eligible to vote, but who have not registered.
The lack of a single term comparable to population is only part of the problem of the open-ended nature of what plaintiffs want. The least open-ended and most readily ascertainable of the possible voter-related bases is probably registered voters. In theory, all that a state would have to do is to go to its voter rolls and count the numbers. However, state voter rolls are notoriously inaccurate because people move (inside the state and out), people die, mistakes are made and not corrected -- despite the mandate of the 2002 federal law requiring states to have up- to-date voter rolls.
Of the remaining possibilities, the "potential voter" category suggested by plaintiffs is the most subjective, most subject to manipulation and most difficult to count and assign a location. Presumably, plaintiffs would start with the total population and then exclude those too young to vote, those who are not U.S. citizens, and those under a felony disability. It is unclear whether the census data has the first two categories broken down by geography to the degree necessary to use in a potential voter system.
Moreover, because the census does not track felons under a disability, another source for that information would have to be found, an especially daunting task since a person may have been convicted in State A, but now be living in State B. Then what about students who are eligible under state law to vote when they are in college in that state, but who may choose to vote in the state where their parents reside? Could they all be excluded because they are not "potential voters," or would the State have to ask each of them their likely preference for voting in the next election, a time-consuming and highly subjective inquiry? In addition to students, there are many young people who move to major metropolitan areas for work, but who may still prefer to vote back home. If determining exact numbers and locations for all properly registered voters is difficult, making that determination for all "potential voters" would be a nightmare.
As for using the numbers of actual voters, it is not hard to determine who has voted in a given election. However, there are major swings in the number of voters even between presidential and mid-term congressional elections, and even greater differences between those and regular off-year elections. If the state has its choice among these, that would further amplify the discretion available to state officials who could choose an election based on whether the turnout favored one party or the other -- precisely what the re-districting cases are designed to prevent, even if they are less than fully effective to stop all gerrymandering.
To assure that states doing re-apportionment follow the law, instead of the political predilections of the officials in charge, there must be one right answer to the question of what the base line should be for drawing districts. The answer that has been consistently used has been population, which is not subject to manipulation like the alternatives. It is based on the federal census that is conducted on the same basis, at the same time and using the same methods, across the United States. That result is compelled for the U.S. House of Representatives by Article I, section 3, and its specific references to the "actual Enumeration" (the census held every ten years) as the basis for allocating seats in the House. Because the one person, one vote requirement is constitutionally required for both the House and all state legislative districts, that strongly suggests that they should have the same base line (denominator) of total population. It would make little sense from a constitutional perspective to have a different base for state elections than for the House, let alone for each state to be permitted to decide for itself which base to use. One person, one vote is a constitutional principle that does not apply differently in different states or at different times. Because the House districts must use population as their base line (not voters or anything else), the Equal Protection doctrine applicable to state districting decisions must also use population if equality is to be assured in practice as well as in theory.