In this latest Supreme Court case Evenwel v. Abbott, exploring the claim that political redistricting should focus on balancing the number of eligible voters in the population across districts as opposed to total population, several amicus briefs have been submitted by interested parties. None other than Nathaniel Persily; Stanford Law professor and a sought after expert on redistricting has submitted this brief along with several other equally accomplished colleagues in the voting/redistricting field: Bernard Grofman, Stephen Ansolabehere, Charles Stewart III, and Bruce Cain. Oral argument is scheduled for December 8th.
The crux of the “Persily” brief is that irregardless of whether the Constitution requires or allows redistricting on the basis of the “eligible” voter population, there is no practical way to date to approximate that population, or in other words, there is no reliable, acceptable dataset that would reveal the number of eligible voters in any location in the U.S.
The brief outlines several possible datasets suggested by the plaintiffs in this case and discusses the rather signifigant shortcomings of each.
Voting Age Population
At the outset, the brief points out that no state nor the federal government maintains a list of “eligible voters,” but the appellants in this case offer possible proxies for rebalancing eligible voters among districts. The First is VAP (voting age population); it is not nearly the equivalent of the “voting eligible” population however. VAP is just that; a count of how many persons in a jurisdiction are of age to vote (18 yrs or older).
The Census Bureau supplies these numbers but these totals do not include members of the military and other overseas workers that are indeed eligible to vote. Second, and even more significant is the fact that VAP numbers include many individuals who have either lost or never acquired the right to vote, such as prisoners, the mentally disabled, disenfranchised felons and noncitizens.
Citizen Voting Age Population
CVAP data can be found in the Census Bureau’s American Community Survey (ACS) dataset, which is based on a survey of just 2.5% of U.S. households. 1year, 3 year and 5 year ACS estimates are published on a rolling basis. Professionals that work with the ACS dataset however, are fully aware of its limitations; mainly that the margin of error for these estimates is considerably high.
The Census Bureau issues its own warning about making some specific assumptions from the data based on that high margin. Persily and his colleagues also point out the unavailability of this CVAP data at the appropriate level of detail. That is, the most detailed level of reporting these statistics is at the census block group level (roughly equivalent to several urban neighborhoods), which generally represents a rather large number of people. Redistricting is routinely performed at the much smaller census block level- roughly equivalent to one or two city blocks.
Lastly, the brief details the inherent weaknesses of voter registration rolls, which are by definition under inclusive of eligible, non-registered individuals of voting age. Persily and company assert that voter registration lists historically fluctuate wildly depending on whether it is an election year, are chronically inaccurate, and the constitutional clearly does not require such a select population base for redistricting.