It is the amicus brief of Nate Persily, Bernie Grofman, Charles Stewart, Steve Ansolabehere, and Bruce Cain.  Hard to think of a better A-list of political scientists working at the intersection of election law and political science.  Here is the summary of the argument:

The United States Constitution requires the creation of a single population dataset: the decennial Census’s “actual enumeration” of persons. U.S. Const. art. I, § 2; amend. XIV. As such, states and localities, almost without exception, have used this dataset to build redistricting plans, and courts have repeatedly upheld plans that do so. Neither the federal government, nor any state, maintains an address list of eligible voters that would allow for redistricting on that basis. Surveys, funded by congressional whim, that provide partial estimates of eligibility based on citizenship are a poor substitute for the census redistricting dataset. An interpretation of the Fourteenth Amendment that would prohibit the use of the most accurate and only constitutionally mandated population dataset and, in effect, mandate the creation of some new count of eligible voters would be both unprecedented and incredibly destabilizing to the U.S. Census and redistricting process.

The contested philosophical arguments occupying most of the briefing in this case can be avoided in favor of a simpler resolution based on the type of population data available and usable for redistricting. Appellants’ interpretation of the constitutional requirement of one person, one vote is radical not only in its theoretical underpinnings, but also in its real, practical implications for the redistricting process. They argue that the dataset all states used for redistricting in 2010 is constitutionally deficient and impermissible. Instead, some other data – perhaps the American Community Survey, registered voter statistics, or some heretofore nonexistent dataset of eligible voters – should be used as the population basis for redistricting. None of these datasets, however, have the granularity, timeliness, detail, or accuracy comparable to the census enumeration.

Appellants’ constitutional argument is predicated on the notion that it is possible to draw districts around equal numbers of eligible voters. If the Fourteenth Amendment requires that only people who can vote should be equally represented, then redistricting, under this view, should be based on equal numbers of eligible voters and no one else. For most states, that means the census enumeration of the total population, plus voting eligible military and overseas voters not counted at their voting address in the census, minus children, noncitizens, prisoners and disenfranchised felons, and those ineligible because of mental disability. No state maintains a dataset of eligible voters, as such.

Appellants, therefore, urge this Court to mandate, as a constitutional rule, the use of currently available second-best alternatives that would not satisfy the rigid legal standard they proffer. Estimates of the citizen voting age population (CVAP) derived from the yearly American Community Survey (ACS) of 2.5 percent of households do not provide current, accurate data at the levels of geography (census block level or precinct) where most redistricting is conducted. At best, the ACS five-year averages give ballpark estimates of previous citizenship rates, several years before redistricting is conducted. The ACS could also be eliminated by the government at any time, as the House of Representatives has voted to do, or fully or partially defunded, as has happened twice since its inception.

Registered voter lists invite a different set of problems, and can only be used for redistricting if they match up well with more reliable population statistics. They are ripe for political manipulation and highly variable depending on the temporal proximity of the list to a given election. Moreover, at least one state does not keep a voter registration list, and another dozen allow for Election Day registration, which can lead to substantial changes in voter registration data in a short period of time.

The one-person, one-vote rule is not broken, and this Court should not try to fix it. The collateral damage caused by a rejection of the census as the basis for redistricting cannot be easily contained. In the end, Appellants not only invite the Court to read the Constitution to prohibit what is now the near-universal use of census population data for redistrict ing, but they also suggest that the Fourteenth Amendment requires government collection of data on voter eligibility that heretofore has not existed. The Court should reject that invitation.