Skip navigation

How Best to Mitigate Prison Gerrymandering

Download original document:
Brief thumbnail
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
1
How Best to do Prison Population Data Reallocation, i.e., How Best to Mitigate Prison
Gerrymandering?*

Bernard Grofman
Distinguished Professor of Political Science
University of California, Irvine
Bgrofman@uci.edu
August 10, 2021

KEY WORDS: prison gerrymandering; one person, one vote; apportionment;
redistricting

* Grofman is Distinguished Professor of Political Science, and Jack W. Peltason Chair,
University of California, Irvine. This research was supported by the Peltason Chair.

Electronic copy available at: https://ssrn.com/abstract=3906274

2
ABSTRACT
We review the legal and practical complexities in doing prisoner population data
reallocation and discuss differences in the practices in different states and jurisdictions.
We emphasize that there are two key complications. (1) how to handle prison
populations that do not have an in-state past address, and (2) potential legal differences
between how prison populations are to be treated for state legislative (or local)
redistricting as opposed to congressional redistricting. We propose a best practices rule.

Electronic copy available at: https://ssrn.com/abstract=3906274

3
As noted on the redistricting website of the National Conference of State Legislatures:
“11 states have passed laws modifying how incarcerated persons are counted and
allocated during the redistricting process, 1 although Illinois' law goes into effect for the
2030 cycle. 2 States vary in their reallocation treatment of state and federal inmates and
in their specific procedures for identifying and reporting incarcerated persons’ last
known residences for redistricting purposes.”

3

The main goal of this essay is to lay out the factual and legal background of debates
about prisoner population data reallocation schemes since I believe that this issue is
likely to be a topic of continuing controversy because (a) the few cases that have been
litigated on the issue of prisoner population data reallocation do not discuss the fact
that, now, different states /localities make different choices about how to implement
prisoner population data reallocation; (b) litigants have not yet called attention
to potential inter-state issues about persons with no legislative representation
anywhere arising when states exclude some out-of-state (out of jurisdiction) prisoners
from any state (local) legislative count, and (c) perhaps most importantly, how to handle
the redistricting base for the U.S House is still in legal limbo, and choice of the
redistricting base for U.S House elections raises unique constitutional and statutory
issues.

A secondary aim of this essay is to provide advice to political jurisdictions and
commissions considering prisoner population data reallocation by identifying the likely
consequences of different rules, though I also provide my own advice on best practices.

Electronic copy available at: https://ssrn.com/abstract=3906274

4
-- while recognizing that more than one practice is likely to be legal, at least for
purposes of legislative redistricting. I see the issue with prisoner data reallocation in the
context of redistricting less whether or not to do it, but rather how to do it.

In the states that have opted for prisoner population data reallocation, the general rule is
that prisoners with former addresses that are in-state are simply assigned to their
previous home address when that is administratively feasible. 4 However, there appear
to be two key complexities -- involving both normative questions and legal ones.

The first of these how do we best deal with the prisoners who either have out-of-state
former residences or whose former residences cannot be located. Colorado and Virginia
assign prisoners with out-of-state addresses or unknown addresses to the prison
address, 5 while Connecticut counts such prisoners for purposes of redistricting as
“generic residents of the state -- as overseas residents and military personnel are
counted.” Other states simply exclude this category of prisoners with unknown or out of
state former addresses from the count used for purposes of districting. 6 In
mathematical terms there is no difference between Connecticut’s assignment of the
residual category to the state as a whole and simply excluding them, since in both
cases the state population for purposes of redistricting will have to exclude this residual
category in order to make the district population totals match the state total. 7 But
Connecticut, unlike the states that simply exclude this residual category, does recognize
that those in this residual category are, for certain limited purposes, to be treated as
residing in the state.

Electronic copy available at: https://ssrn.com/abstract=3906274

5

The second issue is whether or not there are differences between what is legally
permissible vis-à-vis state legislative or local redistricting and what might be mandated
in the case of redistricting for the U.S. House of Representatives.

I am very sympathetic to reallocating prisoners to their previous home addresses for
redistricting purposes since there are many arguments in its favor and the most
plausible legitimate argument against doing so involves administrative convenience. 8
Moreover, the Supreme Court has, in a relatively little noticed 2012 case, upheld the
constitutionality of such reallocations. 9 However, I support a position for both legislative
and congressional redistricting that will make neither opponents of prisoner population
data reallocation nor advocates of an end to “prison gerrymandering” happy, namely, to
use the rule for prisoner population data reallocation used by Colorado and Virginia.
This rule assigns all those whose in-state addresses can be identified to the area of
their previous address but assigns those with out-of-state former addresses, or
unknown addresses, to the prison location. I recognize that the rule used by Virginia
and Colorado continues to provide some overrepresentation to areas (disproportionately
rural) where there are prisons but I see this rule as based on a reasonable tradeoff
among competing representation principles. Moreover, for the U.S. House of
Representatives I suggest reasons why this version of prisoner population data
reallocation might be the only one that can pass constitutional muster.

Electronic copy available at: https://ssrn.com/abstract=3906274

6
The three basic good government arguments for counting prisoners at their previous
home address are straightforward:

10

First, counting individuals for redistricting data purposes at their previous home location
is a more faithful representation of living patterns in the state since imprisonment
(except for a handful of those serving life sentences) is not permanent.

Second, failing to use previous home addresses distorts the representation of
geographic areas, since prisons are not randomly scattered around the state, and
prisoners are disproportionately located in rural areas as compared to the proportion of
prisoners whose previous home addresses were rural. Thus, districts with prisons (and
thus, disproportionately rural areas) are favored vis-a-vis population equality
requirements, and, because felons currently in prison are disenfranchised in almost all
states, that allows election wins with fewer votes in districts with prisons. 11

Third, the allocation of prisoners to rural areas can impact not just urban representation
but also minority representation. Indeed, many of the reformers arguing for prisoner
reallocation have largely been concerned with its racial aspect. Prison population is
disproportionately minority as compared to the total voting age population. 12

Counting

Black population in the prisons where they are incarcerated not only underweights
areas where minority opportunity districts might realistically be drawn, but it can also
lead to an illusion of a Black population majority that is completely ineffective in insuring

Electronic copy available at: https://ssrn.com/abstract=3906274

7
the election of candidates of choice of the African-American community because too
much of the Black voting age population consists of disenfranchised felons. 13

A fourth argument is that representation of prisoner interests will be more likely if their
residence is treated as their past residence. The degree to which representatives from
the areas where prisons are located pay attention to prisoners requires empirical
investigation, with limited attention likely to be especially true for prisoners with out-ofstate past addresses and/or with no voting rights, but some representatives with prisons
in their district might well see prisoners simply as people to be served. For prisoners
with past in-state addresses, if they came to be counted at their prior residence, it is
possible that elected officials who represent the area where the prisoner once lived will
see the prisoner as a potential future constituent as well as a past constituent and this
may further enhance a sense of connection with someone now in prison. 14 Such a
prospect would be still further enhanced if prisoners have family where they formerly
lived.

But I should also note that there are political consequences to not reallocating prisoners
to their previous home addresses, since failing to do so favors areas which are
disproportionately Republican in their voting. Indeed, there is a clear partisan division of
support for a change to prisoner reallocation to their previous home for redistricting
purposes, with Republican legislators largely in opposition and the states which have
decided on some form of prisoner reallocation disproportionately ones controlled by

Electronic copy available at: https://ssrn.com/abstract=3906274

8
Democrats. But the magnitude of such partisan effects requires state specific
evidence. 15

Now I turn to a more detailed look at some of the complexities in doing prisoner
population data reallocation.

First, because of the different ways the various states have chosen to deal with prisoner
data for inmates whose home address is in a different state (or whose home address
cannot be identified), in seven states we have a conflict between the total population
used to allocate seats to the state and the population the state is using for internal
redistricting purposes. In those states (California, Delaware, Illinois, Maryland, New
Jersey, New York, and Washington) prisoners in state facilities with out-of-state
previous addresses are simply excluded from the population count for state legislative
redistricting purposes and, in five states, are excluded for congressional districting
purposes as well. Using total population numbers for redistricting that are different from
those reported to the state by the Census could present a potential state law issue, but
whether it would depends heavily upon the wording of the state’s constitution. 16
Moreover, as I read Burns v. Richardson, 384 U.S. 73 (1966), it says that states are not
required by federal law to use the federal census as the population basis of state
legislative districting. 17

Second, the legal issues in excluding some prisoners from any redistricting population
count seem especially problematic in five states in that the exclusion rules in those

Electronic copy available at: https://ssrn.com/abstract=3906274

9
states do raise potential federal law issues vis-à-vis prisoner reallocation, not just state
law issues, because they exclude prisoners with out-of-state or unknown addresses
from the count not just in state legislative elections but also for elections to Congress. In
California, Illinois, Maryland, and Washington the exclusion of state or federal prisoners
with an out-of-state previous address extends to congressional districts as well as
legislative districts. In New York, according to the NCSL website, all federal prisoners,
regardless of previous home address are excluded from both state level and
congressional redistricting. The exclusion of individuals from congressional population
counts raises several legal issues.

1. Why should a state be allowed to count those excluded individuals in the census
allocation of congressional districts to the state, and then proceed to disregard them in
the state’s own congressional districting? How can you justify counting them for
apportioning districts but excluding them when it comes time to draw districts? 18 I do
not believe that this issue, as opposed to the general question of a state’s right to do its
own legislative redistricting, has ever been squarely posed to a federal court. 19

2. Especially for congressional redistricting, how do we deal with the fact that excluding
data for some prisoners creates a class of “stateless persons.” They are not counted in
the prisons where they are incarcerated, and thus not counted for redistricting purposes
in the states where those prisons are located, but they are also not counted anywhere
else (unless their home address state tries to track them down, but that hasn’t
happened -- and it would not make sense for a state to pursue this once congressional

Electronic copy available at: https://ssrn.com/abstract=3906274

10
apportionment was complete). Thus, the rule of excluding prisoners with out-of-state
former addresses or prisoners with no known previous address entirely from the
redistricting base might still invite litigation, since there is nowhere the interests of such
prisoners would be represented in any way at the legislative level. 20 Moreover, the
incarcerated who cannot be reassigned to the district in the state where they previously
resided because they do not have a known previous home address, or they have an
out-of-state previous home address, is non-trivial in number. 21 In some jurisdictions
they might easily be a quarter or more of the state/federal prison population. 22

3. In Connecticut, the assignment of prisoners with former out-of-state addresses or
missing addresses is to a category of statewide assignment. This might seem to
contradict the federal statute that prohibits at-large congressional districts though it also
might not, since there may be no one in that category who is a voter. The Connecticut
approach to dealing with out-of-state and unassignable prison population has been
defended as optimal by analogy with how military overseas personal are treated, i.e., by
assigning them to a state as a whole rather than to a particular residence within a state
when this is unknown, since overseas personnel are asked to designate a state as a
“home of record” address. 23 I do not find this analogy compelling. -24 Prisoners with no
known previous in-state address are still physically living in a known location within a
state, namely the prison. They are not physically residing outside the U.S.

Third, there is differential treatment of the data for prisoners in state facilities and in
federal facilities in some states, with New York having the starkest difference. 25 In New

Electronic copy available at: https://ssrn.com/abstract=3906274

11
York, while those incarcerated in state facilities with an in-state previous address have
their residence data reallocated to that address, if they are resident in a federal facility,
as noted earlier, all prisoners are excluded from any redistricting population counts
regardless of whether their previous address was in-state or out-of-state. However, this
difference, too, has some further complexity in that a reason that states treat federal
prisoners differently is because the federal Bureau of Prisons has taken the position that
it is banned from sharing information on prisoner former addresses for reasons of
privacy. 26

Fourth, we have the problem of potential legal precedents. If it is legally OK for states to
(a) exclude some persons (in prison) from the state population count used for
redistricting purposes because they are not state residents, and/or (b) use different
population counts for state redistricting and congressional redistricting, what does that
say about the legality of states excluding some or all non-citizens from redistricting
population counts -- at least for state legislative redistricting? In seven states (mostly
Democratic controlled states), the fact that prisoner re-allocation excluded prisoners
with out-of-state home addresses from the state’s redistricting population count may
well be used by conservatives/Republicans to justify exclusion of non-citizens from the
population tallies used for redistricting – at least for state level redistricting.

While I do think that this kind of legal spillover could be a potentially relevant
consideration, since it already seems inevitable that some Republican trifecta states
where state legislatures do the redistricting will opt for excluding non-citizens from the

Electronic copy available at: https://ssrn.com/abstract=3906274

12
redistricting count (and perhaps even propose excluding those below voting age,
though that would seem to run up against the Burns v. Richardson precedent), I doubt
that precedents about how prison population is dealt will be that consequential in how
the Supreme Court ultimately deals with the non-citizens/CVAP issue vis-à-vis choice of
redistricting base (or apportionment base). 27 I think that a far more relevant factor in
how the Supreme Court will approach this issue in the 2020 round will be the feasibility
of identifying an “appropriate” (sic!) class of non-citizens to be excluded. Here there are
multiple problems. First and foremost, there is no information on the 2020 census form
on who is and who is not a citizen. Second, the American Community Survey (ACS),
which does ask a citizenship question, was not conducted in 2020. Third, even if we had
the most recent ACS data, it is based on a sample and is not accurate at the block level
or at precinct (voting tabulation unit) level critical in most redistricting. 28 Fourth,
because the Census publicly reports its data in such a way as to preserve privacy of
information for individuals residing in small census units, matching administrative
records to census data is not that straightforward. Fifth, even if we attempt to use
administrative records to directly determine the size of the non-citizen population, before
we can exclude anyone from the count, we would need to know that they were in the
count in the first place! But there is no way to know who did or did not fill out a census
form without violating the privacy of census respondents. 29

Taking into account the issues discussed above, I favor prisoner population reallocation,
and my proposed “solution” to the “missing persons” problem of prisoner reassignment
for those for whom previous addresses are unavailable or who reside out of state is to

Electronic copy available at: https://ssrn.com/abstract=3906274

13
balance out the competing representational concerns by following the lead of the states
who dealt with the residual category of unassigned prisoners by assigning them to the
prison location. 30

In my view, it is unfortunate that the states who incorporated a policy of exclusion of
some prisoners from the state’s redistricting base in their rules for prisoner reallocation
had not fully considered the consequences of this policy of exclusion of the residual
category from any count. The Colorado and Virginia rules that I regard as best practices
will not make some advocates of an end to prison gerrymandering fully happy even
though they do assign all prisoners with known in-state addresses to their former
address when this is administratively feasible, because unassignable and out-of-state
prisoners will still be counted at the facility, and thus overrepresentation of areas where
there are prisons will not be fully curtailed. Nonetheless, I view this rule as definitely
better than the alternative of not doing prisoner reallocation at all. So, regardless of
whether or not prisoner population data reallocation can be legally compelled,
jurisdictions making choices should, in my view, choose to do prisoner population data
reallocation following the Colorado and Virginia precedents, rather than those from
other states. 31

A NOTE ON TERMINOLOGY: Throughout this essay I use “redistricting” when I am referencing the
population used for line drawing of political subunit boundaries and “reapportionment” to refer to the new
population numbers assigned by the Census to a given state (or other subunit) after each decennial
Census. But this distinction is not universal, with some courts and commissions using apportionment and
districting synonymously. Also, I deliberately use the quite clumsy term “prisoner population data
reallocation” for what is popularly called “prison gerrymandering” and what the National Conference of
State Legislatures refers to as “reallocating inmate data.” Though I use the terms “prisoner
gerrymandering” when I reference essays by prisoner population data reallocation advocates, since that
is their preferred term, I prefer to avoid it since it has such politically laden affect. I chose to use
“prisoners” rather than “inmates” since I see potential differences for legal purposes between prisoners
1

Electronic copy available at: https://ssrn.com/abstract=3906274

14

and other types of institutionalized individuals. Of course, it is important to recognize that not all those
who are listed as residing in the same census block as a prison are incarcerated inmates of that prison.
I am indebted to Rick Pildes, Nick Stephanopoulos, Nate Persily, Gerry Hebert, Wendy Underhill, Ben
Williams, Dana Paikowsky and Rory Kramer for helpful comments on earlier versions of this essay. The
inspiration for writing this essay came from hearing a joint presentation to the Pennsylvania Legislative
Reapportionment Commission of Professor Rory Kramer and Professor Briana Remster, Department of
Sociology and Criminology, Villanova University on the topic of the consequences of prisoner population
data reallocation in Pennsylvania. https://www.redistricting.state.pa.us/resources/video/080421 LRC.mp4.
I am indebted to Jonathan Cervas, Mapping Consultant to the Pennsylvania Reapportionment
Commission, for sharing with me a number of the many excellent presentations to the PRC, including the
one by Kramer and Remster. None of the individuals mentioned above bear any responsibility either for
remaining factual mistakes in this essay or for the point of view it expresses. I have not previously written
about prisoner population data reallocation and there may well be factual or legal errors remaining.
Please feel free to contact me about this essay at BGTravel@uci.edu with corrections or comments.
2

3 Hundreds of county and municipal governments across the country have also implemented rules to
avoid what reform advocates refer to as “ prison gerrymandering.” (Local Governments that Avoid PrisonBased Gerrymandering, Prison Policy Initiative (Jan. 7, 2019),
https://www.prisonersofthecensus.org/local/.) I am indebted to Dana Paikowsky for calling this reference
to my attention.

For those for whom valid street addresses for former residences are available, computerized geocoding techniques exist to reasonably efficiently identify where they should be counted if they are not
counted at the prison. In particular, the Census has the capacity to do so upon request from a state or
other jurisdiction. https://www.census.gov/programs-surveys/decennial-census/about/rdo/summaryfiles/2020/GQAssistance_CensusGeocoder.html We should note, however, that lack of a past street
address may be especially problematic for some minority groups, e.g., native Americans living on tribal
lands where the only address is a post-office box.
4

As of the time of this writing (August 11, 2021), because the Colorado Supreme Court ruled that the
decision about prisoner data reallocation is constitutionally in the hands of the state’s two independent
redistricting commissions rather than the state legislature, Colorado’s rules are actually not yet set. The
commissions are scheduled to reach early decisions on this issue. See
https://coloradonewsline.com/2021/08/10/colorado-redistricting-prison-census-data/
Also, Dana Paikowsky (personal communication, August 11, 2021) indicated that there were at least
three other states with commissions considering the issue of prisoner population data reallocation and
their legal authority to do something about it.
5

As of the date of my NCSL website access (August 6, 2021), Nevada seems not to have addressed this
question.

6

7

Several of those who commented on an earlier draft of this essay made this point.

8

The Census received 77, 795 comments on its proposed 2020 guidelines for identifying residence, of
which 77,887 pertained to prisoners and 44 pertained to overseas military personnel. Of the 77,887
comments pertaining to prison populations, 77,863 suggested that prisoners should be counted at their
home or pre-incarceration address. Federal Register :: Final 2020 Census Residence Criteria and
Residence Situations As summarized in this Census document, the arguments offered in these
comments parallel those given below. I am indebted to Jonathan Cervas for calling this document to my
attention.
In a per curiam opinion in Fletcher v. Lamone, 567 U.S. 930, 133 S. Ct. 29 (2012) the U.S. Supreme
Court upheld Maryland’s 2010 “No Representation Without Population Act,” which counts incarcerated
people at home for redistricting purposes. See Davis, Michelle. 2012. "Assessing the Constitutionality of
Adjusting Prisoner Census Data in Congressional Redistricting: Maryland's Test Case," University of
9

Electronic copy available at: https://ssrn.com/abstract=3906274

15

Baltimore Law Forum 43(1). Although the Supreme Court Opinion was per curiam, the Fletcher
district court opinion Fletcher v. Lamone, 831 F.Supp.2d 887, 894-97 (D. Md. 2011) (three-judge panel)
provided more guidance. It emphasized that "a State may choose to adjust the census data, so long as
those adjustments are thoroughly documented and applied in a nonarbitrary fashion, and they otherwise
do not violate the Constitution." Id. at 894-95. As far as I am aware Fletcher v. Lamone is the only case
directly on this topic which has reached the Supreme Court. However, there have been post-Lamone
cases on prisoner reallocation decided by federal courts. Davidson et al v. City of Cranston, Rhode Island
837 F.3d 135 (U.S. Court of Appeals, First Circuit, September 21, 2016 ) held that the City of Cranston
could not be compelled to reallocate prison population. That court stated: “The Constitution does not
require Cranston to exclude the ACI inmates from its apportionment process, and it gives the federal
courts no power to interfere with Cranston's decision to include them.” (at p. 146). On the other hand,
Calvin et al. v. Jefferson County Board of Commissioners, Florida, 172 F.Supp.3d 1292 (United States
District Court, N.D. Florida, Tallahassee Division, March 19, 2016) found, on equal protection grounds,
that the sizeable prison population district in the county needed to be excluded from the count of persons
used for redistricting purposes in the County Commission because it would make up such a large share
of the district population in any district into which it would be placed. However, that court so emphasized
the peculiarities of the factual context of the case that it was clear that its decision was intended to have
virtually no precedential value. The court noted: “importantly, we are dealing with state prisoners and
a county government. It is the interplay of the limited powers of the county government, the fact that the
prisoners are under state control, and the fact that the prisoners are confined that deprives the prisoners
of a meaningful representational nexus with the county government.” The court also asserted that “the
size of the prison population relative to the size of the districts is such that counting the prisoners makes a
substantial difference. This would not be the case in counties with larger populations.” If we were not
dealing with prisoners or if either of these features were not present, “this would be a different case. In
particular, the situation would be quite different if we were dealing with a state legislative district, because
state prisoners are obviously affected by the policies put in place at the state level. When Mr. Boyd
received letters from JCI inmates, he put those letters aside because there was nothing he could do for
them in his capacity as a County Commissioner. The same would not be true of the state senator and
representative whose territory includes JCI.”
.
10The need for prisoner population data reallocation was decisively raised in Wagner, Peter. 2002.
Importing Constituents: Prisoners and Political Clout in New York: A Prison Policy Initiative Report, April 2
(updated May 20, 2002) https://www.prisonpolicy.org/importing/importing.html .The group he is
associated with was instrumental in pushing this reform idea forward. The idea of prisoner reallocation
began attracting more attention from the voting rights community after three states (Maryland, Delaware,
and New York) adopted a prisoner data reallocation rule in 2010, though in not all these states did this
rule immediately affect redistricting. Interest in prisoner population data reallocation was further raised
with the publication of Persily, Nate. 2010-11. “The Law of the Census: How to Count, What to Count,
Whom to Count, and Where to Count Them,” 32 Cardozo Law Review , 755-790, which highlighted this
issue as a potentially important one for fairness of representation. A very useful reference that provides
relevant data for all 50 states on the consequences of failing to do prisoner reallocation is Prison-based
gerrymandering after the 2010 Census | Prison Gerrymandering Project (prisonersofthecensus.org) . I am
indebted to Dana Paikowsky for calling my attention to this reference.
However, though prison populations are disproportionately rural, we must be careful not to paint this as
purely an urban versus rural issue. There are some prisons in urban areas, and not all rural areas are
equally advantaged in population terms by counting prisoners at the prison, since there are many districts
with no prisons. Also, some rural areas without a prison may gain representation from reallocating
prisoners from proximate rural counties that do have prisons. (
For relevant calculations for the State of Pennsylvania, see also their August 4. 2021 joint presentation to
the Pennsylvania Legislative Reapportionment Commission.
https://www.redistricting.state.pa.us/resources/video/080421 LRC.mp4 ) See also Remster, Brian and
Rory Kramer. 2018. “Shifting Power: The Impact of Incarceration on Political Power.” Du Bois Review,
15(2): 417–439.
11

Electronic copy available at: https://ssrn.com/abstract=3906274

16

12See e.g., discussion pp. 1-7, Texas Civil Rights Project, Prison Gerrymandering Report, 2021.
https://txcivilrights.org/wp-content/uploads/2021/04/Prison_Gerrymandering_Report.pdf

A leading advocate of prisoner reallocation, Peter Wagner (2012. “Beginning of the end for ‘prisonbased gerrymandering’.” Washington Post, July 13) gives an illustration from a local government in
Maryland to show that this is not merely a hypothetical possibility. And the same problem can arise with
Hispanics and other groups. When I have served as a court-appointed Special Master, in examining the
potential for geographically compact “minority opportunity districts” that respect existing political subunit
boundaries, to avoid vote dilution I have tried to be careful to retabulate the “effective majorities” for
minority groups in districts where there are prisons so as to take into account only those prisoners
(minority and non-minority) who actually could participate in the political process.
13

Especially for minorities, this connection may be enhanced if the representative from the previous
address is of the same race or ethnicity as the prisoner (and whether or not this is true for the
representative of the district that includes the prison). The political science literature has long noted the
phenomenon of some representatives of the same ethnicity as an individual seeing themselves as, at
least symbolically, as representatives of those with similar racial/ethnic background.
14

See also previous endnote that discusses the degree to which counting prisoners at their prison
provides excessive representation to rural areas.
15

There are important differences across state constitutions. For example, the New York State
Constitution (Article 3, Section 4) requires use of Census data only "insofar as such census and the
tabulation thereof purport to give the information necessary therefore." But the Alaska Const. Art. VI, §
3 reads: “The Redistricting Board shall reapportion the house of representatives and the senate
immediately following the official reporting of each decennial census of the United States.
Reapportionment shall be based upon the population within each house and senate district as reported
by the official decennial census of the United States.” See
https://www.ncsl.org/research/redistricting/redistricting-and-use-of-census-data.aspx for other states. I am
indebted to Wendy Underhill for the reference to the Alaska Constitution and to this part of the NCSL
website.
16

17 However, in Burns, when considering Hawaii’s use of registered voters as a potential redistricting base,
the Court essentially said that this is a prohibited unless you can demonstrate that you get essentially the
same results as you get from using "state citizen population." Thus, if state citizen population is taken to
be limited to state residents, then Burns seems to imply that excluding prisoners with out-of-state former
addresses from state redistricting may not be a problem in terms of federal law. However, as Persily
(2010: 775) points out, subsequent to Burns, in Karcher v. Daggett, 462 U.S. 725, 731 (1983) the Court
opined that "[a]dopting any standard other than population equality, using the best census data available
... would subtly erode the Constitution's ideal of equal representation" (emphasis added by Persily). Also
Burns involved state legislative redistricting not elections for the U.S. House. How, in the light of Evenwel
v. Abbott, 136 S. Ct. 1120 (2015), the present Supreme Court will respond to any challenges to particular
rules for prisoner reallocation or exclusion, and which, if any rule, can be legally mandated by one person,
one vote or equal protection concerns is, in my view, still very much an open question and might depend
upon case-specific facts, and on whether or not the elections are state or local level or Congressional.
Like other election law scholars such as Nicholas Stephanopoulos (personal communication, August 8,
2021) I do not see a contradiction between the holdings in Burns and in Evenwel vis-a-vis state
legislative population bases. Burns says there are circumstances where it is possible for states to choose
to use apportionment bases other than total population in legislative districting, while Evenwel says that
states are not required to use CVAP as the population base for districting.

Here I pose this a potential legal problem but acknowledge that different rules govern inter-state data
as opposed to intra-state redistricting.
18

19

But, as a non-lawyer, I may well have missed relevant cases in my search.

Electronic copy available at: https://ssrn.com/abstract=3906274

17

20 Here the potential litigants would be different from those in earlier lawsuits, i.e., they would be out-ofstate prisoners excluded from representation asking to be counted somewhere so that there might be
some representatives to whom they might appeal who might regard them, however faintly, as
constituents.

There are potential parallels with military base population and with university students residing in
college dormitories. But there are also key differences. In particular, for federal elections, on-base military
personnel in the U.S. cannot be denied the right to vote in federal elections in the state in which the base
is located, but the Census will use administrative records to assure that military overseas personnel can
receive mailed ballots appropriate for the federal elections at their designated home address. The Census
counts domestically based military personnel at their base location if they were there on April 1, 2020,
though that date may have been affected by the restart of the Census in response to the COVID-19
pandemic. See https://www.census.gov/content/dam/Census/library/factsheets/2020/dec/2020-how-wecount-military.pdf and Counting All Military Service Members and Their Families in 2020 (census.gov) In
most states, college students residing in dorms can choose where to have themselves registered, either
at a prior residence or in the dorm, but for population assignment purposes the Census uses their
address as where they were found during the census count in 2020. This means that the COVID-19
pandemic could have affected some student locations since most college dorms were closed during the
early months of the pandemic. The trial court in Lamone rejected the argument that it was improper for a
state to adjust census data to account for prisoners without also adjusting for college students and
members of the military. It asserted (831 F.Supp.2d at 896) that "college students and military personnel
have the liberty to interact with members of the surrounding community and to engage fully in civic life. In
this … sense, both groups have a much more substantial connection to, and effect on, the communities
where they reside than do prisoners." Id. But that court also noted "prisoners are counted [by the Census
Bureau] where they are incarcerated for pragmatic and administrative reasons, not legal ones." Id. at 895.
(I am indebted to Rick Pildes for reminding me about these analogies, and to Nate Persily for identifying
relevant references.) For the 2020 redistricting the Census solicited public input about how various
groups such as students in dorms, those is facilities for the aged, prisoners, etc. should be treated on the
census for residence purposes.
21

But there are also questions about whether all those with previous addresses should be assigned to
those addresses. For example, should someone with a life sentence not be assigned to his or her prison
where they will be spending the rest of their life, rather than to a previous address. And what about
someone whose sentence would keep him or her in prison until past the next reapportionment period?
These are, however, minor and essentially technical questions which are best resolved in the legislation
that sets up the rules for prisoner reallocation in states that choose to do so.
22

See Justin Levitt. “Address Unknown” — Podcast Episode #1 | Prison Gerrymandering Project
(prisonersofthecensus.org) I am indebted to Dana Paikowsky for this reference.
23

24 The constitutionality of this rule for dealing with overseas military personnel was upheld in Franklin v.
Massacusetts 505 U.S. 788 (1992). I am indebted to Dana Paikowsky for calling this case to my attention.

Generally, prisoners with out-of-state former addresses are most likely in federal prison, next most
likely in state prisons, and least likely in county jails.
25

26I

agree with the point made by voting rights advocates that implementation of prisoner reallocation
requires full cooperation by the Federal Bureau of Prisons in providing information about prisoners’ past
addresses. See e.g., “ Letter to DOJ on Routine Usage Exception to Allow for Implementation of
Universal Enfranchisement and Abolition of Prison Gerrymandering,” from Campaign Legal Center,
League of Women Voters of the United States and Washington Lawyers’ Committee for Civil Rights and
Urban Affairs to Attorney General Merrick Garland, March 30, 2021. Letter to DOJ on Routine Usage
Exception to Allow for Implementation of Universal Enfranchisement and Abolition of Prison
Gerrymandering | Campaign Legal Center. I am indebted to Dana Paikowsky for calling this reference to
my attention.

Electronic copy available at: https://ssrn.com/abstract=3906274

18

It is interesting that a trial court considering the issue of excluding prison population within a County
Board of Commissioners’ district in which a prison was located (Calvin et al. v. Jefferson County Board
of Commissioners, Florida, 172 F.Supp.3d 1292 (United States District Court, N.D. Florida, Tallahassee
Division, March 19, 2016) considered but dismissed possible spill over consequences for decisions about
other forms of exclusion, such as exclusion of non-citizens or minors. “Defendants have predictably
made a ‘slippery slope’ argument. They warn that "should th[is] Court enter the arena of determining
which individuals are worthy of being included in an entity's population data, it should not be unexpected
that arguments to exclude other segments of the population will shortly follow. Policy arguments exist to
exclude resident aliens and minors, or to give more weight to areas that have a high concentration of
eligible voters, such as areas with high concentrations of the elderly as opposed to younger families with
children. This Court should not wade into these political judgments. … I am convinced that the slope
ahead is not so slippery. …. Prisoners are not like minors, or resident aliens, or children — they are
separated from the rest of society and mostly unable to participate in civic life.”
27

See “Brief in the U.S. Supreme Court of Nathaniel Persily, Kenneth Prewitt, Bernard Grofman, Stephen
Ansolabehere, Charles Stewart III, and Bruce Cain as Amici Curiae in Support of Appellees” (filed
September 18, 2015) in Evenwel v. Abbott, 136 S. Ct. 112 (2016).
28

Note that I am not arguing that it will never be possible for states to use CVAP as an apportionment
base for legislative redistricting. That is a legal question about which my crystal ball is fogged over. I do
think it is quite possible that, ultimately, the Supreme Court will deal with the issue of an appropriate
population base for redistricting in much the same way as it did in one person, one vote cases, i.e., by
speaking with a forked tongue. Just as it distinguished the standard for state level redistricting vis-a-vis
“one person, one vote” from the OPOV standards for states or local redistricting, it could proceed
analogously with respect to the differential treatment of data on prison inmate population for federal as
opposed to state or local redistricting purposes. But, in 2022, as I indicated above, deleting just those
non-citizens who did actually fill out a census form presents special and almost certainly insurmountable
technical problems worsened by the time crunch for 2021-22 redistricting caused by Census data
collecting and reporting delays attributable to COVID-19 (and possibly also to underfunding of this
Census). Nonetheless, I anticipate that, looking ahead to 2024, some states will be so insistent on not
counting non-citizens for apportionment purposes that they will conduct state censuses of their own that
do include a citizenship question and seek to use that data to do mid-decadal legislative redistricting. This
may not violate state constitutional provisions in some states (see earlier footnote discussing important
differences across state constitutions in the language referencing the census and NCSL reference cited
therein) but may well raise federal “equal protection” issues. Moreover, the reliability of using such nonCensus based counts, or using ACS data that is only from a population sample and is not accurate down
to the block or precinct level, can be and will be challenged (see e.g.,
https://www.yalelawjournal.org/forum/beyond-the-adjustment-wars ).
29

30 Counting only prisoners with no in-state previous addresses at the prison has the pragmatic justification
that, if state legislative district population variance is (barring specific forms of intentional and
unconstitutional population-based gerrymandering) acceptable within a 10% total deviation rule it is
unlikely that there will be sufficiently many “residual” prisoners counted as living at the prison for there to
be a violation of one person, one vote if they were removed entirely from the count. This point is made is
the presentation to the Pennsylvania Legislative Reapportionment Commission of Professor Rory
Kramer and Professor Briana Remster, Department of Sociology and Criminology referenced in a
precious footnote This possibility of such a “hypothetical” one person, one vote violation can be still
further reduced, in a different kind of “compromise” between the seeming dichotomy of doing prisoner
reallocation and not doing prisoner reallocation, by requiring that districts that contain a prison not be
underpopulated relative to the ideal district population--though there is a question about the
constitutionality of such an approach in the light of Cox v. Larios 542 U.S. 947 (2004). I should note that
the idea of such a “compromise” is not original with me.

In the best of all worlds one might say that prisoners should get reallocated to their former home state
and their old address, rather than being “captured” (to use the term used by some reform advocates) in a
federal prison in a different state, but doing this simultaneously to change all the population numbers in
31

Electronic copy available at: https://ssrn.com/abstract=3906274

19

all the states is administratively infeasible because different states have different rules and may be
changing those rules post-census, and/or finish the reallocation task at different points in time. In any
case, the inter-state coordination burden seems immense.

Electronic copy available at: https://ssrn.com/abstract=3906274