Has Congress been improperly counting prisoners in the Census since 1790?

Following up on yesterday's take on the Rhode Island "prison gerrymandering" case, a few more thoughts come to mind.

The district court linked apportionment to redistricting (as the Supreme Court in Evenwel did), identifying women, children, slaves, and other non-voters as appropriately included within the apportionment base. The court went on

The inmates . . . share none of the characteristics of the constituencies described by the Supreme Court. They don't have a stake in the Cranston public school system and they are not receiving constituent services, such as help with public-benefits bureaucracies. They are not making requests of and suggestions to Cranston elected officials (or if they are, they are receiving no response), nor are they receiving "the protection of government," at least not from Cranston elected officials.

The Census Bureau has, since 1790, counted prisoners as residents of the place where they are incarcerated. When Congress apportions members of Congress to the states, it includes those prison populations--even if a number of the prisoners come from out-of-state.

Does this mean Congress has been unconstitutionally apportioning members of Congress since 1790? That prisoners should have been counted as residents of their home states for the last 220 years, and Congress has simply got it wrong this whole time?

Surely, the swift rejoinder to this claim is that very few long-term prisoners were included in a Census, much less out-of-state prisoners, until recently. But this was the same claim raised by the plaintiffs in Evenwel--that the United States had a very small non-citizen population until recently, a claim that was soundly rejected.

Further, the analogy of apportionment to redistricting grows even worse when one considers redistricting itself--where should prisoners be included for redistricting purposes?

Suppose one makes the claim that prisoners should be located as residents of the address they resided at before being incarcerated. Does this somehow cure the concerns raised by the district court? Suppose you are an incarcerated prisoner in Cranston who was originally residing in Providence. Do you suddenly "have a stake in the Providence public school system"? Do you "receiv[e] constituent services"?

The dangerously overbroad language from the district court would effectively make prisoners invisible! If they are constitutionally forbidden from being considered members of the political community where they are incarcerated, surely considering them members of the political community where they last resided is even more attenuated. It risks making prisoners invisible, uncounted members of any political community.

Instead, the better claim--and it is not a terribly good constitutional claim at that--is that this artificial concentration of non-voters into a single location is suboptimal for other voters in adjacent districts, and an attempt should be made to more naturally redistribute those non-voters in some way.

But this is not a constitutional claim. It is simply a way--perhaps a better way--of drawing districts. But the Supreme Court's decision in the "one person, one vote" cases imposed a one-size-fits-all requirement to redistricting. And rhetoric in cases like this Rhode Island prison gerrymandering case further complicate any meaningful understanding of what that mantra means.