The fundamental nature of the right to vote was under scrutiny during the 2018 midterm elections, involving ongoing allegations in several states of voter suppression, eligibility and fraud. Another aspect of the right to vote was on the ballot in Florida, where voters approved a referendum that lifted the lifelong voting ban imposed on otherwise eligible voters who had been convicted of a felony. This column addresses yet another facet of the right to vote: its relative strength, as it is affected by a practice colloquially known as “prison gerrymandering” through which incarcerated people are counted as residents of the towns where they are imprisoned (rather than where they lived before they were incarcerated) for purposes of drawing parameters for legislative districts.

New York and Maryland are the only states that expressly outlaw the practice of prison gerrymandering. Instead, they require that incarcerated persons be counted as residents of their pre-incarceration addresses for representation purposes. The New York and Maryland “anti-prison gerrymandering” laws recognize that inmates are not functionally residents of the towns where their prisons are located, that they do not freely live in those towns, and if they were at liberty, they would reside where they did before they were imprisoned.

The constitutionality of “prison gerrymandering” was examined by two courts in 2016: the Northern District of Florida (Jefferson County) and the U.S. Court of Appeals for the First Circuit (City of Cranston). Notably, these courts came to opposite conclusions with Jefferson County holding the practice unconstitutional and City of Cranston finding the practice constitutional. This summer, the NAACP filed a new challenge to the constitutionality of prison gerrymandering in the District of Connecticut (NAACP v. Merrill).

The Merrill case may result in Second Circuit and Supreme Court review on this divided issue. As in the two previous cases, the complaint alleges that the practice violates the “one person, one vote” requirements of the Fourteenth Amendment by significantly increasing the number of residents in certain state districts without increasing the number of actual constituents in those districts, which gives voters in those districts greater voting power than voters in other districts. Plaintiffs allege that the problem is even more severe in Connecticut, where the communities surrounding prison facilities are populated by groups with different racial and socioeconomic makeups than the inmates' home communities. According to the Merrill complaint, there are twice as many African Americans and Latinos as there are white people in Connecticut prisons. While a large number of African Americans and Latino inmates maintained pre-incarceration addresses in one of Connecticut's three urban centers (Hartford, Bridgeport, and New Haven), the districts in which the prison facilities sit are almost entirely rural, predominantly white, lightly-populated, areas.

'Prison Gerrymandering' in Context

After the Florida referendum, convicted felons are now entitled to vote in the great majority of states (excepting Iowa, Kentucky and Virginia), typically after release from incarceration (as in New York, Connecticut and New Jersey) or upon completion of parole/probation. Only two states (Maine and Vermont) allow incarcerated persons to vote while in prison. A small minority of states have additional requirements, in part related to the underlying crime, that remain significant barriers to voting for convicted felons. State approaches to franchise disenfranchisement vary tremendously. See generally Felony Disenfranchisement Laws (Map), ACLU (last visited Dec. 26, 2018).

The Supreme Court has interpreted the Equal Protection Clause of the Fourteenth Amendment to require that states abide by the “one-person, one-vote principle,” which requires states to establish congressional districts with populations that are as close to equal as possible. In assessing what is “equal” under that principle, the Supreme Court has established a “safe harbor” rule for state and local districts, which presumes that a districting plan is presumptively compliant with the constitution's one-person, one-vote principle if the maximum deviation between the most and least populated voting districts is less than ten percent. While deviations that exceed the 10 percent “safe harbor” are not per se unconstitutional, they are presumptively impermissible and shift the burden to the challenged jurisdiction to show legitimate interests justifying the deviation. See Wesberry v. Sanders, 376 U.S. 1, 7-8 (1964) (holding that congressional districts must be drawn so that “as nearly as is practicable one man's vote in a congressional election is worth as much as another's”); Reynolds v. Sims, 377 U.S. 533, 568 (1964) (holding that electoral maps must achieve “substantial equality of population among the various districts” and that deviations from population equality between districts can only be justified if “based on legitimate considerations incident to the effectuation of a rational state policy”); Evenwel v. Abbott, 136 S. Ct. 1120, 1124 (2016) (“[w]here the maximum population deviation between the largest and smallest district is less than 10%, the Court has held, a state or local legislative map presumptively complies with the one-person, one-vote rule”).

Congressional jurisdictions typically use the population as determined by the constitutionally mandated census as the base population for apportioning voting districts. In reaching that figure, the census uses the “usual residence” rule to determine where people live, which it defines as “the place where a person lives and sleeps most of the time.” See U.S. Census Bureau, Residence Rule and Residence Situations for the 2020 Census (last visited Dec. 18, 2018) (hereinafter Residence Rules and Situations for the 2020 Census). While courts have interpreted the “usual residence” rule to mean “more than mere physical presence” and has emphasized that individuals should possess “some element of allegiance or enduring tie” (see Franklin v. Massachusetts, 505 U.S. 788, 789 (1992); Residence Rules and Situations for the 2020 Census, para. 15) when assessing whether certain groups should be counted as members of a community in which they temporarily reside, incarcerated persons continue to be counted in 48 states as residents of the district in which they are incarcerated as opposed to their pre-incarceration (and likely post-incarceration) residences. Incarcerated persons thus become “ghost constituents” to whom elected representatives have no accountability, but whose presence contributes to the existence of their districts.

Is 'Prison Gerrymandering' Unconstitutional?

As noted above, there is conflicting precedent on the constitutionality of prison gerrymandering. In Calvin v. Jefferson County Board of Commissioners, the Northern District of Florida held in 2016 that counting incarcerated persons who could not vote as part of the baseline population for determining voting districts violated the Equal Protection Clause of the Fourteenth Amendment as those residents lacked a “meaningful representational nexus” to their legislative body. In reaching this conclusion, the court noted that (1) the incarcerated persons at issue were isolated from the surrounding community; (2) the representatives at issue possessed little legal authority or practical ability to substantially affect the incarcerated persons' lives through their policies; and (3) the representatives at issue made very little, if any, effort to solicit the input of the incarcerated persons at issue. See Calvin v. Jefferson County Board of Commissioners, 172 F. Supp. 3d 1292, 1321 (N.D. Fla. 2016).

In contrast, also in 2016, in Davidson v. City of Cranston, Rhode Island, 837 F.3d 135, 142-44 (1st Cir. 2016), the First Circuit reversed the district court's finding that “prison gerrymandering” violated constitutional norms, finding that the practice was constitutional in light of the Supreme Court's 2016 holding in Evenwel v. Abbott, which was decided two weeks after Jefferson County. In Evenwel, the Supreme Court rejected an equal protection challenge to the use of total-population census data when redrawing Texas' state senate districts, finding that the “one-person, one-vote” principle did not require states to use voter population when apportioning voting districts. The Supreme Court found that requiring “voter-eligible apportionment as constitutional command would upset a well-functioning approach to districting that all 50 States and countless local jurisdictions have followed for decades, even centuries,” and observed that nonvoters can constitute a legitimate population for apportionment purposes given that nonvoters have an important stake in many policy debates.

Expanding upon the Supreme Court's refusal to require states to count only the voter population when apportioning districts, the First Circuit held that prison gerrymandering was constitutional. The First Circuit found that “the use of total population from the Census for apportionment is the constitutional default, but certain deviations are permissible, such as the exclusion of non-permanent residents, inmates, or non-citizens immigrants.” The First Circuit continued that these “permissible deviations are just that—optional, but not the norm”—and that Rhode Island's calculations were “plainly permissible.” The First Circuit further asserted that, where districts fell within the “safe harbor” under the one-person, one-vote principle, plaintiffs must present a showing of “invidious discrimination” to call such districts into constitutional question. It concluded that prison gerrymandering alone did not rise to that level.

'NAACP v. Merrill'

NAACP, et al. v. Denise Merrill, Secretary of State and Dannel P. Malloy, Governor, No. 3:18-cv-01094 (N.D. Conn.) is the first known case since City of Cranston to challenge the constitutionality of prison gerrymandering. Like plaintiffs in Jefferson County and City of Cranston, plaintiffs in Merrill argue that prison gerrymandering violates the one-person, one-vote principle of the Equal Protection Clause. Plaintiffs in Merrill further assert that incarcerated persons should properly be counted as “residents” of their pre-incarceration residences (not “residents” of the facilities in which they are imprisoned) because the discrepancies created by Connecticut's use of prison gerrymandering exceed the “safe harbor” under the one-person, one-vote principle. To support that argument, plaintiffs in Merrill highlight the socioeconomic makeup of Connecticut's prison population, specifically noting that a disproportionate number of incarcerated persons are African-American and Latino who are moved from their homes in diverse, urban communities to facilities in rural, white districts when incarcerated.

The state of Connecticut has moved to dismiss the action, alleging that plaintiffs' call for the state to count incarcerated people as residents of their pre-incarceration communities is asking for it to modify census data when measuring population for purposes of establishing voting districts. Defendants argue that, based on the First Circuit's interpretation of Evenwel, Connecticut's method of counting residents and apportioning voting districts is constitutional under the Equal Protection Clause of the Fourteenth Amendment. Defendants further assert that any discrepancies created by prison gerrymandering fit squarely into the “safe harbor” of the one-person, one-vote principle when incarcerated persons are counted as “residents” of their facilities.

In response, plaintiffs argue that “the fundamental principle of representative government in this country … is equal representation for equal numbers of people” and that prisoners are “essentially unrepresented” by the elected officials in the districts where they are incarcerated, and as such, Evenwel does not control. Plaintiffs propose counting incarcerated people as residents in their districts of origin because the elected officials in those districts are more likely to affect their lives. This is based not only on the fact that inmates will likely return to their previous places of residence, but also because inmates' families (including, but not limited to, children who are ineligible to vote) remain there during their sentences. NAACP v. Merrill, No. 3:18-cv-01094 (N.D. Conn.). (ECF Nos. 1, 14, 21).

Conclusion

The District of Connecticut will assess whether a challenge to the constitutionality of prison gerrymandering, under the circumstances alleged in the complaint, remains a legally cognizable claim following the Supreme Court's determination in Evenwel. Additional appellate review is likely to follow no matter which way the court decides. In the wake of the 2018 midterms and the issues surrounding the right to vote, prison gerrymandering is a captivating but little publicized aspect of that debate. For now, incarcerated persons in the 48 states that do not ban prison gerrymandering remain “ghost constituents” silently haunting the voting districts in which their prisons sit and disproportionately empowering their neighbors.

Steven M. Witzel is a partner of Fried, Frank, Harris, Shriver & Jacobson and chair of the white-collar defense, regulatory enforcement and investigations practice. Amanda Giglio is an associate at the firm.