II. Federal Law Does Not Require Reliance on the Census Bureau

While state and local governments are required by federal law to redistrict each decade and typically use the decennial census to do so, federal law does not require that choice. Most governments rely on the U.S. Census for redistricting because the data is high quality and free. But the Supreme Court has said that states, and local governments by extension, are free to use other sources of data.

One Supreme Court case, Burns v. Richardson, implicitly approved the type of adjustments for prison populations discussed here:

Neither in Reynolds v. Sims nor in any other decision has this Court suggested that the States are required to include . . . persons denied the vote for conviction of crime in the apportionment base by which their legislators are distributed and against which compliance with the Equal Protection Clause is to be measured. The decision to include or exclude any such group involves choices about the nature of representation with which we have been shown no constitutionally founded reason to interfere.

When states draw their congressional districts, they are required to use the best population data available, and, although as a general matter, Census data will be the best available, states are not required to use data they know to be flawed, just because it is Census data.

Similarly, the Supreme Court in Mahan v. Howell rejected Virginia's argument that it was compelled to use Census Bureau assignments of residences of military personnel in its state legislative redistricting and suggested that a state may not use Census data it knows to be incorrect.

State and local governments are therefore free, at least under federal law, to create their own censuses from scratch, or to simply correct how the federal Census counts people in prison. Recently, a federal three-judge panel specifically rejected claims that adjusting Census data to count incarcerated people as residents of their legal home addresses for redistricting purposes is unconstitutional.

Not only does the federal government not require state and local governments to count incarcerated people at the prison location, but the Census Bureau has recently begun helping these governments do the opposite. The Census Bureau recently changed how it publishes its data to make it easier for states and municipalities to draw districts without including the prison population. Although this is not commonly understood, there is no one monolithic decennial Census. That is, there are actually at least three different decennial data products provided by the U.S. Census Bureau that are relevant to redistricting, each with its own purpose, methodology, and numbers. The first is the state-level counts, which include military and federal employees overseas and is used for Congressional apportionment. The second is the PL94-171 redistricting data file, produced since 1980, that is most commonly used for redistricting. The third is the Advance Group Quarters Summary file, produced for the first time after the 2010 Census, for the explicit purpose of helping governments with the problem of prison-based gerrymandering. As the Census Bureau explains:

This early release of data on the group quarters population may be beneficial to many data users including those in the redistricting community who must consider whether to include or exclude certain populations in redrawing boundaries . . . . It will permit state and local redistricting officials to overlay this file with the 2010 Census Redistricting Data (Public Law 94-171) Summary File data.

This redistricting cycle, Maryland and New York both relied on the Advanced Group Quarters Summary file in conjunction with the state's corrections departments' data to adjust the data used for redistricting. Courts in both states have approved the laws requiring these adjustments in order to count incarcerated people at their home addresses for redistricting purposes.