\\server05\productn\P\PLR\24-2\PLR218.txt

unknown

Seq: 1

16-FEB-05

12:49

Prisoners of the Census: Electoral and
Financial Consequences of Counting
Prisoners Where They Go, Not Where
They Come From*
Eric Lotke†
Peter Wagner‡
Mass incarceration distorts society in peculiar ways. Obviously there are individual effects, felt keenly by the 2.1 million
people who wake up behind bars every morning.1 There are social effects, experienced by the approximately two million children with a parent in custody2 and by neighborhoods where
whole segments of the population have been removed.3 Lastly,
* This research was supported by grants from the Soros Justice Fellowship
Program of the Open Society Institute.
† Eric Lotke is Director of Research and Policy at the Justice Policy Institute.
He performed this research as a Soros Senior Justice Fellow. Previously, he was
the Executive Director of the D.C. Prisoners Legal Services Project, a judicial clerk
on the Supreme Court of Connecticut, and an adjunct professor at Georgetown and
George Washington Law Schools. He is a graduate of Wesleyan University and of
the University of Wisconsin.
‡ Peter Wagner, JD, is a Soros Justice Fellow and Assistant Director of the
Prison Policy Initiative, a widely-used internet project providing accurate, timely
research and policy reports on criminal justice issues. He is the author of Importing Constituents: Prisoners and Political Clout in New York, the first systematic
state analysis of the impact of prisoner enumeration policies on legislative redistricting. As part of his Soros Justice Fellowship, he is currently measuring the
impact of prisoner enumeration on legislative redistricting in a number of states
and publishing the results on the project’s website, PrisonersoftheCensus.org.
1. BUREAU OF JUSTICE STATISTICS, U.S. DEP’T OF JUSTICE, BULL. NO. NCJ
200248, PRISONERS IN 2002 1 (2003), available at http://www.ojp.usdoj.gov/bjs/pub/
pdf/p02.pdf [hereinafter PRISONERS IN 2002].
2. See BUREAU OF JUSTICE STATISTICS, U.S. DEP’T OF JUSTICE, BULL. NO. NCJ
175688, WOMEN OFFENDERS 8 tbl.18 (1999), available at http://www.ojp.usdoj.gov/
bjs/pub/pdf/wo.pdf; see also BUREAU OF JUSTICE STATISTICS, U.S. DEP’T OF JUSTICE,
BULL. NO. NCJ 182335, INCARCERATED PARENTS AND THEIR CHILDREN (2000),
available at http://www.ojp.usdoj.gov/bjs/pub/pdf/iptc.pdf.
3. See, e.g., Todd Clear, The Problem with “Addition by Subtraction”, in INVISIBLE PUNISHMENT: THE COLLATERAL CONSEQUENCES OF MASS IMPRISONMENT 181
(Marc Mauer & Meda Chesney-Lind eds., 2002).

587

\\server05\productn\P\PLR\24-2\PLR218.txt

588

unknown

PACE LAW REVIEW

Seq: 2

16-FEB-05

12:49

[Vol. 24:587

there are budgetary effects such as states cutting education
budgets while prison budgets continue at record highs.4
Other effects of mass incarceration are so subtle they pass
without notice. Basic tools of democratic society slip out of
place and cease to function properly. The census is one of those
tools.
Obtaining an accurate count of the population is so fundamental to representative democracy that the framers of the
Constitution required it in the opening paragraphs. Article I
Section 2 of the U.S. Constitution requires an “actual enumeration” of the population every ten years “in such manner as they
[the Congress] shall by law direct.”5 This enumeration is used
to apportion voting representation, draw political boundaries
and allocate funds among state and local governments. Taking
the 2000 decennial census required “the largest peacetime mobilization in the nation’s history . . . .”6
But mass incarceration distorts this fundamental tool. The
U.S. Census Bureau counts people in prison where their bodies
are confined—in prison—not the communities they come from
and where they are genuine members.7 This would be an item
of statistical trivia, but the new numbers give it new meaning.
More people now live in prison and jail than in our three least
populous states combined.8 Organized differently, they would
4. See, e.g., JUSTICE POLICY INST, Cellblocks or Classrooms?: The Funding of
Higher Education and Corrections and It’s Impact on African American Men
(2002), available at http://www.justicepolicy.org/article.php?list=type&type=20.
5. U.S. CONST. art. I., § 2; Congress fulfills this function in the Census Act, 13
U.S.C. § 141 (1976).
6. Press Release, U.S. Census Bureau, Census Workers Ready to Contact
Households That Did Not Respond to Census 2000 (Apr. 25, 2000), available at
http://www.census.gov/Press-Release/www/2000/cb00cn37.html.
7. See U.S. CENSUS BUREAU, U.S. DEP’T OF COMMERCE, 2000 CENSUS OF POPULATION AND HOUSING: SUMMARY FILE 3 TECHNICAL DOCUMENTATION C-2 (2002),
available at http://web.archive.org/web/20030405051219/http://www.census.gov/
prod/cen2000/doc/sf3.pdf [hereinafter SUMMARY FILE 3].
8. According to the Census, the combined population of the three smallest
states is 1,729,541 (Wyoming, 493,782; Vermont, 608,827; and Alaska, 626,931).
U.S. Census Bureau, Population Estimates, at http://eire.census.gov/popest/data/
states/files/ST-EST2003-AS200004.csv (Apr. 1, 2000). According to PRISONERS IN
2002, the total number people in confinement are 2,166,260 (prisons, 1,440,655;
local jails, 665,475; juvenile, military, immigration and other facilities, 139,527).
PRISONERS IN 2002, supra note 1, at 1.

\\server05\productn\P\PLR\24-2\PLR218.txt

2004]

unknown

Seq: 3

PRISONERS OF THE CENSUS

16-FEB-05

12:49

589

have six votes in the United States Senate. It is not trivia
anymore.
Yet organization is precisely the problem. The high rates of
incarceration are not evenly distributed across the population.
White men are imprisoned at a rate of 912 per 100,000; black
men are imprisoned at a rate of 3,437 per 100,000.9 Moreover,
imprisonment moves people in predictable patterns—typically
out of large urban centers and into rural communities.
Whether these differences reflect different involvement in criminal behavior or selective enforcement is actually beside the
point. From the point of view of the Census Bureau, it does not
matter. If people are in prison, that’s where their bodies count.
This article discusses the primary consequences of the way
the Census Bureau counts people in prison—the impact on electoral apportionment and financial distributions. It maps the
U.S. population, explains how and why the Census Bureau acts
as it does, and suggests possible reforms. In brief, the article
finds consistent, low-level distortions in both voting and funding that could be avoided if the Census Bureau counted people
differently.
I. The Rules
The first step is understanding how the census works. The
Census Bureau’s general rule is to count people in their “usual
residence,” the place where they live and sleep most of the time.
The usual residence need not be the same as a person’s legal or
voting address, and a person need not be there on the literal
census day (April 1st).10 They can take a vacation and still
count at home.
Determining the usual residence for most people is easy.
However, special categories present special challenges. Sailors
in the merchant marine, children in joint custody and long-term
commuters all require special rules, and these rules have
evolved over time.11 People in prison are in a category called
9. PRISONERS IN 2002, supra note 1, at 9 tbl.14.
10. U.S. Census Bureau, Facts About Census 2000 Residence Rules: The Concept of Unusual Residence, at http://www.census.gov/population/www/censusdata/
resid_rules.html (last visited Mar. 20, 2004) [hereinafter Residence Rules].
11. Peter Wagner, Usual Residence Rule Has Been Modified for Other Special
Populations and Can be Changed for Prisoners Too, PrisonersoftheCensus.org, at

\\server05\productn\P\PLR\24-2\PLR218.txt

590

unknown

PACE LAW REVIEW

Seq: 4

16-FEB-05

12:49

[Vol. 24:587

“group quarters” which includes nursing homes, college dormitories, military installations and other places where unrelated
persons live together.12 As a rule, people in group quarters
count where the group quarters are located.13 For people in
prison, that’s the prison.
Discretionary decisions are made about whether people in
prison will be given census forms to fill out themselves, or
whether the warden will simply provide a headcount. One in
six people in prison is given a “long form” to fill out with additional information, just like in the general population.14 However, if the person provides anything other than the
institutional address, that information is discarded.
II. The Map
The best way to see how prisons move people is to create a
map. Fully 5% of all growth in the U.S. rural population in the
1980’s was people in prison.15 In the 1990’s, an astonishing 30%
of new residents of upstate New York were brought there
against their will.16 Guard towers are slowly replacing small
towns and family farms as the struggling heartland turns to
prisons as an industry of last resort.
All kinds of communities are affected. West Feliciana Parish, Louisiana, is classified by the Census Bureau as a 100%
rural community, but 5,000 of its 15,000 residents live in custody—fully a third.17 In comparison, Walker County, Texas is
http://www.prisonersofthecensus.org/news/fact_of_the_week-archive-3-11-2003.
shtml (Nov. 3, 2003).
12. Residence Rules, supra note 10, at Question 11.
13. Id.
14. Interview with Edison Gore, U.S. Census Bureau, Assistant Division
Chief for Planning Decennial Management Division (May 19, 2003).
15. Calvin L. Beale, Prisons, Population, and Jobs in Nonmetro America, 8
RURAL DEV. PERSP’S 16, 17 (1993).
16. ROLF PENDALL, UPSTATE NEW YORK’S POPULATION PLATEAU: THE THIRDSLOWEST ‘STATE’ (2003), available at http://www.brookings.edu/dybdocroot/es/urban/publications/200308_Pendall.pdf.
17. Compare U.S. CENSUS BUREAU, GROUP QUARTERS POPULATION BY GROUP
QUARTERS TYPE, at http://factfinder.census.gov/servlet/DTTable?_bm=Y&-state=
DT&-context=DT&-ds_name=DEC_2000_SF1_U&-mt_name=DEC_2000_SF1_U_
PCT016&-tree_id=4001&-all_geo_types=N&-geo_id=05000US22125&-search_results=01000US&-format=&-_lang=EN (last visited Apr. 14, 2004), with U.S. CENSUS BUREAU, STATE AND COUNTY QUICKFACTS, WEST FELICIANA PARISH, LOUISIANA,

\\server05\productn\P\PLR\24-2\PLR218.txt

2004]

unknown

Seq: 5

16-FEB-05

PRISONERS OF THE CENSUS

12:49

591

more urban, with small cities totaling 60,000 people—but more
than 13,000 of them cannot eat breakfast without permission of
the warden.18 Altogether, nearly 200 counties in America have
more than 5% of their population in prison.19 Eighteen counties
have more than 20% of their population in prison.20 This extraordinary transformation can be seen in the fine print of the
2000 census but it is unnoticed unless looked for.21
Top Twenty Prison Counties22
State
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

Louisiana
Texas
Florida
Illinois
Tennessee
Virginia
Texas
California
Texas
Missouri
Texas
Texas
Texas
Texas
Arkansas
Texas
Illinois
Nevada
Texas
Virginia

County
W. Feliciana Parish
Concho
Union
Brown
Lake
Greensville
Mitchell
Lassen
Hartley
DeKalb
Jones
Walker
Bee
Childress
Lincoln
Madison
Johnson
Pershing
Anderson
Sussex

Population

% Rural

Prison
Pop.

Prison
%

15,111
3,966
13,442
6,950
7,954
11,560
9,698
33,828
5,537
11,597
20,785
61,758
32,359
7,688
14,492
12,940
12,878
6,693
55,109
12,504

100.0%
100.0%
52.2%
41.8%
100.0%
64.4%
32.0%
58.7%
57.6%
67.1%
60.9%
36.3%
30.6%
34.0%
100.0%
69.9%
79.1%
100.0%
41.3%
100.0%

4,995
1,299
4,052
1,912
2,090
3,027
2,523
8,367
1,343
2,626
4,650
13,691
7,070
1,652
3,003
2,681
2,640
1,370
10,750
2,379

33.1%
32.8%
30.1%
27.5%
26.3%
26.2%
26.0%
24.7%
24.3%
22.6%
22.4%
22.2%
21.8%
21.5%
20.7%
20.7%
20.5%
20.5%
19.5%
19.0%

at http://quickfacts.census.gov/qfd/states/22/22125.html (last visited Apr. 14,
2004).
18. Compare U.S. CENSUS BUREAU, GROUP QUARTERS POPULATION BY GROUP
QUARTERS TYPE, at http://factfinder.census.gov/servlet/DTTable?_bm=Y&-state=
DT&-context=DT&-ds_name=DEC_2000_SF1_U&-mt_name=DEC_2000_SF1_U_
PCT016&-tree_id=4001&-redoLog=TR]ue&-all_geo_types=N&-geo_id=05000US4
8471&-search_results=01000US&-format=&-_lang=EN (last visited Apr. 14,
2004), with U.S. CENSUS BUREAU, STATE AND COUNTY QUICKFACTS, WALKER
COUNTY, TEXAS, at http://quickfacts.census.gov/qfd/states/48/48471.html (last visited Apr. 14, 2004).
19. One hundred and ninety-seven counties out of 3140, or 6.3% of all counties
have more than 5% of their population in prison. See generally U.S. CENSUS BUREAU, U.S. DEP’T OF COMMERCE, 2000 CENSUS, at http://www.census.gov/ (specific
data sets are on file with the author) [hereinafter 2000 CENSUS].
20. See id.
21. See U.S. CENSUS BUREAU, U.S. DEP’T OF COMMERCE, 2000 CENSUS OF POPULATION AND HOUSING: SUMMARY FILE 1 TECHNICAL DOCUMENTATION 6-68 to 6-69
(2002), available at http://www.census.gov/prod/cen2000/doc/sf1.pdf.
22. See 2000 CENSUS, supra note 19.

\\server05\productn\P\PLR\24-2\PLR218.txt

592

unknown

PACE LAW REVIEW

Seq: 6

16-FEB-05

12:49

[Vol. 24:587

Because no distinction is made between being in or out of
custody, people in prison count the same as shoppers in the local markets, parents whose kids will attend local schools, or
people whose political interests are represented by the person
who represents them in the legislature. Yet these individuals
are not generally from the county where the census has them
placed. They were imported from other counties for purposes of
confinement. If the doors were opened, few would stay.
If the doors were open, likely they would return to where
they came from. That’s the place that most people in prison
consider their home, and where most will return within a few
years.23 The Census Bureau makes no effort to track point of
origin, but geographical information about convictions is generally available from state departments of corrections or the state
judicial branch. Comparing the two data sets reveals where
people are coming from and where they are going to.
Texas, for example, hosts nine of the top twenty counties
ranked by the percentage of population in prison. None of these
counties, however, convicts substantial numbers of people. The
table below shows how people are shipped from one county to
another. The top set of counties are all exporters and the bottom are importers. Dallas, for example, convicts 15.1% of the
people in Texas state prisons but confines none of them. In contrast, Walker County confines 10.4% of the state prison population but convicts hardly anybody.

23. The average time served in state prisons is 3 years. See BUREAU OF JUSSTATISTICS, U.S. DEP’T OF JUSTICE, BULL. NO. NCJ 198821, FELONY
SENTENCES IN STATE COURTS, 2000 3 (2003), available at http://www.ojp.usdoj.gov/
bjs/pub/pdf/fssc00.pdf (last visited Apr. 2, 2004).
TICE

\\server05\productn\P\PLR\24-2\PLR218.txt

2004]

unknown

Seq: 7

16-FEB-05

PRISONERS OF THE CENSUS

12:49

593

Texas Counties that Convict or Confine People24

County
Exporters
Harris (Houston)
Dallas
Tarrant
Bexar
Importers
Walker
Anderson
Brazoria
Coryell
Bee

% State
% State
% County
Prisoners Prisoners Population.
Convicted
Held
In State
Population
There
There
Prison
3,400,578
2,218,899
1,446,219
1,392,931

21.2%
15.1%
7.7%
6.1%

2.1%
0.0%
0.0%
0.0%

0.1%
0.0%
0.0%
0.0%

61,758
55,109
241,767
74,978
32,359

0.2%
0.2%
0.8%
0.1%
0.1%

10.4%
8.2%
6.5%
6.1%
5.4%

22.2%
19.5%
3.6%
10.7%
21.8%

Whatever benefit accrues to a jurisdiction by virtue of its
population, the urban counties are all losing it. Conversely, the
rural counties are getting more than their fair share. The next
sections discuss these benefits.
III. The Vote
The official constitutional purpose of the census is political
apportionment.25 An accurate count of the population ensures
that each state’s delegation in the U.S. House of Representatives and the Electoral College is appropriately proportioned.
The principle of “one person, one vote,” which started as a requirement in the federal system, has since trickled down to
states and even local governments.26 All states base their legislative redistricting on U.S. census data.27
A cursory glance indicates the potential distortion of counting prisons as the “usual residence.” Nearly 9% of all African
American men in their twenties and thirties live in prison.28
Most of this group is apportioned to legislative districts that do
not reflect their communities of interest or their personal politi24. See 2000 CENSUS, supra note 19; see also Texas Department of Criminal
Justice, at http://www.tdcj.state.tx.us.
25. See U.S. CONST. art. I, § 2, cl. 3.
26. See Reynolds v. Sims, 377 U.S. 533 (1964).
27. See, e.g., N.Y. CONST. art. 3, § 4.
28. PRISONERS 2002, supra note 1, at 9 tbl.14 (for ages 20-24 the fraction in
prison is 7.5%; ages 25-29 is 10.4%; ages 30-34 is 8.9%; ages 35-39 is 7.9%).

\\server05\productn\P\PLR\24-2\PLR218.txt

594

unknown

PACE LAW REVIEW

Seq: 8

16-FEB-05

12:49

[Vol. 24:587

cal concerns. Whether they can or do vote is irrelevant; their
bodies still count in the prison district. A more refined analysis
shows that the impact is modest in U.S. Congressional Districts
but more significant in state legislative districts.
The impact of incarceration on apportionment in the U.S.
House of Representatives is small because most people are incarcerated within their own state and because congressional
districts are so large. With districts containing roughly 645,000
people,29 a few thousand people in prison are unlikely to have
much impact. Exceptions may arise as more states send people
out of state, as was the case in 1999 when Wisconsin Representative Mark Green feared that Wisconsin’s plan to export almost 10,000 people to prisons in other states might tip
Wisconsin towards losing a seat in reapportionment.30
The impact within state legislatures, however, is more substantial. The range in size of state legislative chambers is enormous, but the median state Senate district has 106,362
residents and the median state House district has 37,564.31
Thus, just one prison of 1,000 cells is nearly 3% of the population of a median sized House district. Given the frequent clustering of prisons, the impact can accumulate dangerously.
Significant densities of prisoners in legislative districts are especially important because most criminal justice policy is made
in the states.
Under White v. Regester,32 state legislative districts are not
permitted to deviate in size by more than 10%. However, an
analysis of New York State reveals that people in prison put

29. Press Release, Election Data Services, 2000 Census Counts Produce Surprises in Congressional Delegations (Dec. 28, 2000), at http://www.electiondataservices.com/Apport00release_wtables.htm (last visited Apr. 2, 2004).
30. See H.R. 1632, 106th Cong. (1st Sess. 1999) (a bill proposed to provide that
certain attribution rules be applied with respect to the counting of certain prisoners in a decennial census of population); see also Oversight of the 2000 Census:
Examining the Bureau’s Policy to Count Prisoners, Military Personnel, and Americans Residing Overseas, Hearing on H.R. 1632 Before the Subcomm. on the Census
of the House Comm. on Government Reform, 106th Cong (1999).
31. National Conference of State Legislatures, Constituents per State Legislative District: Legislatures Ranked by Size, at http://www.ncsl.org/programs/legman/elect/cnstprst.htm (last visited Apr. 2, 2004).
32. 412 U.S. 755 (1973).

\\server05\productn\P\PLR\24-2\PLR218.txt

2004]

unknown

Seq: 9

PRISONERS OF THE CENSUS

16-FEB-05

12:49

595

district deviations over the maximum range.33 The most significant work to date is Importing Constituents: Prisoners and Political Clout in New York.34 This report documents that in four
New York Senate districts and in ten Assembly districts more
than 2% of the constituents are in prison.35 Analysis in other
states and at the county level is ongoing.

33. PETER WAGNER, IMPORTING CONSTITUENTS: PRISONERS AND POLITICAL
CLOUT IN NEW YORK: A PRISON POLICY INITIATIVE REPORT 10-12 (2002) available at
http://www.prisonpolicy.org/importing/importing_body.pdf [hereinafter IMPORTING
CONSTITUENTS].
34. Id.
35. Id. at 8.

Kirwan
Manning
McDonald
Ortloff
Destito
Nortz
Winner
Nesbitt
Smith
Burling

100
103
112
114
116
118
137
139
146
147

129,732
128,212
129,570
132,349
127,574
128,234
126,784
127,916
131,864
125,572

Reported
Population
3,650
2,793
3,611
9,251
6,187
4,245
3,033
2,851
3,864
6,386

Prisoners
2.8%
2.2%
2.8%
7.0%
4.8%
3.3%
2.4%
2.2%
2.9%
5.1%

% District
Incarcerated
1,934
1,436
1,849
4,623
2,986
2,238
1,665
1,528
1,767
3,269

Black
Prisoners
11.8%
34.1%
96.1%
82.6%
33.5%
68.7%
40.0%
41.3%
56.3%
78.4%

% Black Adults
Disenfranchised

2.5%
1.3%
2.4%
4.6%
0.8%
1.4%
0.2%
1.1%
4.2%
−0.7%

−0.3%
−0.9%
−0.4%
−2.7%
−4.0%
−2.0%
−2.2%
−1.1%
1.2%
−5.8%

Actual
Deviation
from Ideal
District
Size

unknown
Seq: 10

PACE LAW REVIEW

16-FEB-05

36. See PETER WAGNER, IMPORTING CONSTITUENTS: PRISONERS AND POLITICAL CLOUT IN NEW YORK: A PRISON POLICY INITIATIVE
REPORT APPENDIX 35 (2002) available at http://www.prisonpolicy.org/importing/importing_append.pdf [hereinafter IMPORTING
CONSTITUENTS APPENDIX].

AssemblyPerson

Reported
Deviation
from Ideal
District
Size

596

District
#

New York Rural Assembly Districts, 200236

\\server05\productn\P\PLR\24-2\PLR218.txt
12:49

[Vol. 24:587

\\server05\productn\P\PLR\24-2\PLR218.txt

2004]

unknown

Seq: 11

PRISONERS OF THE CENSUS

16-FEB-05

12:49

597

The district of Republican Assemblyperson Chris Ortloff contains the highest percentage of people in prison: 7%.37 The population represented by Assemblyperson Ortloff includes 9,251
prisoners, of whom 4,623 are Black.38 In a district that is already 89% White, 82.6% “of the Black adults in Ortloff’s district
are barred by law from ever voting for or against him.”39 By the
time these prisoners complete their sentences and are again allowed to vote, they will be back home in a different district.
Similar deviations exist in the New York State Senate. All
of the Senate districts in urban Queens County in New York
City were drawn to contain between 12,409 and 12,412 too
many people, a deviation of +4.05%.40 Conversely, a number of
rural Senate districts are short as many as 15,147 residents for
a deviation of -4.95%.41 By the official numbers, the deviation
between districts is 9.78%, slightly less than the maximum allowed.42 But if the prisoners were counted where they actually
are from, the deviation between over-populated Queens and
under-populated rural senate districts would rise to 11.4%,43
more than allowed by White.44 Senator Volker, for example,
represents just 285,305 free people in his rural district; Senator
Maltese from Queens represents 318,484.45 The result is that
each free resident of a rural district with prisons gets a larger
voice in the state capitol than residents of districts in Queens.

37. IMPORTING CONSTITUENTS, supra note 33, at 8.
38. IMPORTING CONSTITUENTS APPENDIX, supra note 36, at 35.
39. See IMPORTING CONSTITUENTS, supra note 33, at 8; see also N.Y. State Leg.
Task Force on Demographic Research and Reapportionment, New York Assembly
District 114 2 (2002), available at http://latfor.state.ny.us/maps/propassem/fa114.
pdf.
40. IMPORTING CONSTITUENTS, supra note 33, at 11.
41. IMPORTING CONSTITUENTS APPENDIX, supra note 36, at 28-30 fig10.
42. IMPORTING CONSTITUENTS, supra note 33, at 10.
43. See id. at 11 figs.3-4 (the difference between the maximum figures is
11.4%).
44. White v. Regester, 412 U.S. 755 (1973).
45. IMPORTING CONSTITUENTS, supra note 33, at 11.

6,623
12,989
5,291
8,951

Prisoners
2.2%
4.3%
4.5%
3.0%

% District
Incarcerated
3,441
6,501
2,791
4,447

Black
Prisoners
19.4%
76.7%
38.9%
75.4%

% Black Adults
Disenfranchised

318,481
318,482
318,484
318,484
318,481
318,484
318,483

1,800
1,800
1,800
1,800
1,800
1,800
1,800

0
0
604
0
0
0
0

Prisoners
Counted at a
Prison Within
District to
Remove

320,281
320,282
319,680
320,284
320,281
320,284
320,283

Estimate of
Actual
Population

4.6%
4.6%
4.6%
4.5%
4.6%
4.6%
4.6%

unknown
Seq: 12

PACE LAW REVIEW

16-FEB-05

46. See IMPORTING CONSTITUENTS APPENDIX, supra note 36.
47. See id.

Senator, July 2002
A. Smith
Padavan
Onorato
(no incumbent)
M. Smith
Maltese
Hevesi & Stavisky

Reported
Population

NYC
Prisoners
Counted
Elsewhere
(estimate)

Actual Deviation
from Ideal
District Size

−3.6%
−6.4%
−6.7%
−6.8%

−1.5%
−2.1%
−4.9%
−3.9%

Urban Queens County (New York City) Senate Districts, 200247

301,528
299,603
290,925
294,256

10
11
12
13
14
15
16

Saland
Stafford
Wright
Volker

41
45
48
59

Reported
Population

Actual
Deviation
from Ideal
District
Size

Reported
Deviation
from Ideal
District
Size

598

Senate
District

Senator

District
#

New York Rural Senate Districts, 200246

\\server05\productn\P\PLR\24-2\PLR218.txt
12:49

[Vol. 24:587

\\server05\productn\P\PLR\24-2\PLR218.txt

2004]

unknown

Seq: 13

PRISONERS OF THE CENSUS

16-FEB-05

12:49

599

It is no coincidence, then, that some of the strongest proponents of incarcerative policies were upstate Republican Senators Volker and Nozzolio, heads of the Committees on Codes
and Crime, respectively.48 Prisons are not just big business in
their districts, they inflate the political clout of every real rural
constituent. Going in to the 2002 redistricting, the prisons in
their two districts held more than 23% of the state’s prisoners.49
Senator Volker readily admits that he does not represent
the prisoners in any real sense. He told Newhouse News Service
that he regularly receives letters from prisoners but that his
real attention is directed toward corrections workers, with
whom he has forged strong relationships. Volker is glad the
prisoners in his district cannot vote, because if they could, “they
wouldn’t vote for me.”50
Overall, in the states, counting urban residents as rural residents dilutes urban voting strength and increases the weight
of a vote in the rural districts. In the rural prison districts, the
real residents benefit because their own issues can receive individual attention from their representative on a scale unavailable elsewhere. In contrast, urban legislators are responsible not
only to their “official” district but also those community members miscounted in the prison diaspora. One can only imagine
the political negotiations of reapportionment, and how a plum
like a prison must count.
Below the state level, in county and town governments, the
impacts become more profound. In Mansfield, Ohio, two large
state prisons make up more than half of Ward 5,51 giving the
voting residents disproportionate voice. In many cases, though,
the results are so obvious, unexpected and unfair that people in
prison are removed from the counts.52 In Greene County, New
48. Editorial, Full-Employment Prisons, N.Y. TIMES, Aug. 23, 2001, at A18.
49. These figures were somewhat lower after the 2000 redistricting, because
the large increase in the prisoner population during the previous decade, forced
these legislators to “share the wealth” with their neighboring but non-prison hosting districts.
50. Jonathan Tilove, Minority Prison Inmates Skew Local Populations as
States Redistrict, NEWHOUSE NEWS SERV., Mar. 12, 2002, at http://www.newhouse
news.com/archive/story1a031202.html.
51. Linda Martz, Taxpayers Before Wards, MANSFIELD NEWS J., Nov. 26, 2002,
at 6A.
52. Tilove, supra note 50.

\\server05\productn\P\PLR\24-2\PLR218.txt

600

unknown

PACE LAW REVIEW

Seq: 14

16-FEB-05

12:49

[Vol. 24:587

York, where the prison town of Coxsackie would have earned
almost another full seat in the county legislature, the legislature voted to remove the prison population from the apportionment entirely.53 Those individuals simply disappear. Of course,
exclusion is logical because people in prison aren’t real residents, but it seems legally and morally inconsistent to retain
them in some apportionment schemes but not others. Similarly, Iberville Parish, Louisiana excluded the prison population
from school board redistricting to avoid drawing a district that
only had two eligible voters.54 The inconsistencies make the
problem obvious: Gardner, Massachusetts was happy to claim
people in prison for state legislative and congressional purposes, but it excluded them in the City Council redistricting
process because the 892 disenfranchised prisoners would have
earned their own district.55 Thus, politicians at different levels
seek to count things in different ways.
The majority of states have constitutional provisions or
statutes defining the important principle of residence for electoral purposes. While most state constitutions authorize the disenfranchisement of people in prison, they also offer explicit
instructions that their residence does not change by virtue of
incarceration. Indeed, the New York Constitution declares: “no
person shall be deemed to have gained or lost a residence, by
reason of his presence or absence . . . while confined in any public prison.”56 The Massachusetts Constitution’s definition of “inhabitant” has led the state supreme court to doubt the state
constitutionality of this application of the “usual residence”
rule, inviting litigation in that state.57
53. Peter Wagner, Prisoners Skew Local Rural Redistricting Too, PrisonersoftheCensus.org, at http://www.prisonersofthecensus.org/news/fact-22-9-2003.
shtml (Sept. 22, 2003).
54. Tilove, supra note 50.
55. Mary Jo Hill, Gardner Excludes Prison Inmates from Political Map,
WORCESTER TELEGRAM AND GAZETTE, June 5, 2001, at B4.
56. N.Y. CONST. art 2, § 4. The New York State Constitution foresees the possibility that the census could be inadequate for state redistricting purposes in that
it requires use of census data only “in so far as such census and the tabulation
thereof purport to give the information necessary therefore” and mandates a special state census to fill in the gaps. N.Y. CONST. art 3, § 4.
57. “We think it clear without elaboration that a census that determines the
place of which a person is an inhabitant on the basis of where he or she lives and
sleeps most of the time will not satisfy the requirement of the Constitution of the
Commonwealth that a person be assigned as an inhabitant to the place of his or

\\server05\productn\P\PLR\24-2\PLR218.txt

2004]

unknown

Seq: 15

PRISONERS OF THE CENSUS

16-FEB-05

12:49

601

IV. The Money
Larger places typically have greater needs and receive a
corresponding share of government resources. Typically they
contribute more as well, though it is not a dollar-for-dollar correlation. One measure of size for determining resource distribution is the official U.S. census population. The involuntary
movement of the population creates a consistent, low-level distortion in funding formulations.
North Carolina distributes up to ¤ cents per dollar of sales
tax to counties and municipalities on the basis of their populations.58 Virginia distributes state aid for K-12 education on the
basis of a complex formula that includes, among other things,
the county population.59 Thus, rural counties that import people for prisons come out ahead of urban counties that send people away. The U.S. Department of Agriculture (USDA)
her domicil.” Opinion of the Justices to the House of Representatives, 365 Mass
661, 663-64 (1974); see also Blanchard v Stearns, 46 Mass 298, 304 (1843); Opinion
of the Justices to the House of Representatives, 122 Mass. 594, 597, 599 (1877).
58. Approximately one half of 11/2 cents per dollar sales tax is distributed per
capita. However, counties can opt in or opt out of some tax collections, and there
are internal distributions within counties. See N.C. GEN STAT. §§ 105-463 to 105520 (2002). The impact is moderated in North Carolina because the prison population is distributed among small county prisons close to the original home. However, the state is considering a move to larger regional prisons. See Dana Damico,
Proposal Would Relieve Crowded Prisons By Easing Sentencing Rules, WINSTON
SALEM J., Apr. 23, 2003, at 1. If it does, the present sales tax structure will create
inequities. The inequalities will be especially ironic because the funds must be
spent on educational capital outlays, so the schools of exporting jurisdictions will
decay but prison towns will improve.
59. The statutory authority is distributed throughout title 22 of the Virginia
code, updated by current appropriations legislation. See, e.g., VA. CODE ANN.
§ 58.1-638 (Michie 2001). Population affects the “composite index.” See JOINT LEGISLATIVE AUDIT AND REVIEW COMM’N OF THE VA. GEN. ASSEMBLY, REVIEW OF ELEMENTARY AND SECONDARY SCHOOL FUNDING (2002), available at http://leg2.state.va.
us/dls/h&sdocs.nsf/By+Year/HD0C2002/$file/rpt277.pdf. The impact of population
can be seen by using the macro-enabled spreadsheets provided by the Virginia Department of Education, and simply adjusting the population variable. See VA.
DEP’T OF EDUC., COMPOSITE INDEX OF LOCAL ABILITY TO PAY (2002-04), available at
index worksheet is at http://www.pen.k12.va.us/VDOE/Finance/Buget/2004cistmplate.xls (last visited Apr. 2, 2004); see also VA. DEP’T OF EDUC., FINAL FY 2003
DIRECT AID ENTITLEMENTS (CHAPTER 1042) BASED ON MARCH 31, 2003 AMD AS OF
JUNE 6, 2003 (STATE FUNDS ONLY), available at http://www.pen.k12.va.us/VDOE/
Finance/Budget/CalcTool-FinalEntitlements.xls (last visited Apr. 2, 2004). The
city of Richmond loses roughly $230,000 annually. The prison county of Sussex,
gains approximately $90,000 annually as a result of its prison population. Virginia is exceptional in using raw population in this way.

\\server05\productn\P\PLR\24-2\PLR218.txt

602

unknown

PACE LAW REVIEW

Seq: 16

16-FEB-05

12:49

[Vol. 24:587

distributes $60 million annually to impoverished Appalachian
communities via the Appalachian Regional Commission.60 Population is a distribution factor, so rural communities with prisons have an advantage over communities without prisons.
There is no indication that the USDA intended to reward prison
construction, but it has that result.
The amount of money at stake is difficult to generalize.
Some states (e.g., Texas) move hardly any revenues between geographic regions or levels of government whereas other states
(e.g., Arizona) move funds generously. Similarly, many specialty taxes (e.g., liquor taxes, cigarette taxes, recreational park
usage fees and hunting-fishing license fees) are distributed on
the basis of population, but a specialty tax that moves significant revenues in one state might not even exist in another.61
The largest funding flows—schools, health and highways—tend
not to depend upon population.
With these cautions firmly in mind, it is still worth drawing
some broad, general conclusions. All things considered, the total impact of counting prisoners in their institutional communities rather than their communities of origin runs a range
between $50 and $250 per person. This estimate is based upon
budgetary analysis of numerous jurisdictions, ranging from
towns to states, as well as news accounts and interviews with
responsible officials. The impact is seldom far below $50 per
head and rarely in the high $200’s. Most frequently it appears
to run in the range of $100 per head.
Thus, when a jurisdiction announces plans to open a new
1,000-bed prison, pause to consider: It will likely generate something like $100,000 in new, “unearned” revenues. Of course, the
estimate could be wrong by over 100% in either direction—depending upon what state it is in, how that state shares reve60. See The Catalog of Fed. Domestic Assistance, 23.001 Appalachian Regional Development, available at http://www.cfda.gov/public/viewprog.asp?progid=
678 (last visited Apr. 2, 2004); see also Appalachian Regional Commission, at www.
arc.gov (last visited Apr. 2, 2004).
61. Ironically, many of these specialty taxes fund programs that exclude by
definition people in prison. For example, the state hunter-fishing license fee might
be distributed for the purpose of improving sports gaming resources, and it might
be distributed to counties on the basis of population, but people in prison will never
use this resource in a county that receives a greater share as a formula artifact of
the prison population.

\\server05\productn\P\PLR\24-2\PLR218.txt

2004]

unknown

Seq: 17

PRISONERS OF THE CENSUS

16-FEB-05

12:49

603

nues, and the particularities of the specialty taxes—but
doubling it or halving it still gives a general idea.62 In percentage terms, $100,000 is seldom a great deal of money, but in dollar terms it always is. Times are tight and towns are counting
every nickel. That $100,000 means a new fire truck, a free renovation for the youth center, or the computer upgrade that was
cut from last year’s budget. When a new wing opens, every additional prison bed will bring an additional $100—or
thereabouts.
The most dramatic impact is in Arizona, a state with significant revenue sharing and large prisons. Florence, Arizona, has
a free population of roughly 5,000 plus another 12,000 living
under lock and key.63 The state and federal funds specifically
linked to the incarcerated population have been estimated at $4
million annually, compared to $1.8 million for the free residents
and $2.3 million raised locally.64 Such lucre tempted the Arizona town of Buckeye to annex nearby Lewis State Prison, population 4,600, though first it had to defeat a matching attempt
by neighboring Gila Bend.65 The mayor of Buckeye, with a population 5,038 before the annexation, promised to use the expected $1.3 million to upgrade parks and family services, and
assured everybody that it would more than pay for the additional burden on fire and police.66
But it is important to note that the new funds do not usually come from the exporting jurisdiction. New York City does
not lose what Attica gains. The export of 43,000 New Yorkers to
62. Researchers who want more detail should focus on taxes on the sale of
liquor and cigarettes, and the lottery. These sources generate sizeable revenues,
exist in most states, and often consider population in the distribution formulas.
However, the largest pots of money—schools, health and highways—typically do
not have population variables. The public officials to call with questions tend to be
in state and county budget or finance offices. Elected officials and professionals in
development or planning offices seldom have precise knowledge of the relevant
formulas.
63. See 2000 CENSUS, supra note 19.
64. Nicholas Kulish, Annexing the Penitentiary, WALL ST. J., Aug. 9, 2001, at
A1.
65. See Beth DeFalco, Buckey Wins Bid to Annex Lewis State Prison, ARIZ.
REPUBLIC, Dec. 22, 1999, at B1; see also S. Comm. on Judiciary, 1999 Leg., 1st Reg.
Sess. (Ariz. 1999) (committee meeting minutes), available at http://www.azleg.
state.az.us/FormatDocument.asp?inDoc=/Legtext/44leg/1R/comm_min/Senate/02
02JUD%2EWPD.htm [hereinafter Arizona Committee on Judiciary].
66. Arizona Committee on Judiciary, supra note 65.

\\server05\productn\P\PLR\24-2\PLR218.txt

604

unknown

PACE LAW REVIEW

Seq: 18

16-FEB-05

12:49

[Vol. 24:587

upstate towns does not affect the gigantic overall population of
New York City. Although the exported New Yorkers come from
a small number of “hotspot” neighborhoods, the budget allocations within the city are not localized in that way. Departments
of sanitation and fire have total budgets, and those budgets are
distributed to districts within the city—but population is not a
variable. Similar analysis leads to similar conclusions in large
cities such as Dallas, Texas, and smaller ones such as New Haven, Connecticut.
Indeed, the most likely losers are similar jurisdictions that
share the same pot of money. For example, the USDA Appalachian Regional Commission grants,67 use population as a variable. However, the urban community that exported those
prisoners isn’t eligible for those funds anyway. It isn’t rural.
The winner is the rural prison town and the loser is the similarly situated community without a prison.
Other funding implications are even more subtle. People in
prison tend to be male and they tend to be members of minority
groups—typically more so than the host community. Thus, the
host town shows odd spikes in gender and racial distribution,
which can improve their prospects in formula grant allocations,
especially programs intended to assist minorities. Furthermore, people in prison are not technically part of the workforce,
and residents of group quarters are not members of households.68 Consequently, they do not count towards unemployment or poverty rates in a community.69 They do, however,
affect per capita income because it is calculated simply by dividing the total community income by the total population.70 Thus,
assistance that targets communities with high unemployment
is not affected but assistance that targets communities with low
per capita income can be distorted. Ron Roth, the planning director of Coxsackie, New York, where prisoners make up 28% of
the town population of 7,600, admitted to Newsday that the
formula depression of per capita income makes the town “more
67. See supra note 60 and accompanying text.
68. SUMMARY FILE 3, supra note 7 (these determinations were made by an
analysis of the technical definitions in collaboration with experts at the Census
Bureau).
69. Id. at 9-6 to 9-6 & n.1
70. Id. at B-20.

\\server05\productn\P\PLR\24-2\PLR218.txt

2004]

unknown

Seq: 19

PRISONERS OF THE CENSUS

16-FEB-05

12:49

605

competitive” for U.S. Housing and Urban Development (HUD)
grants aimed at low income communities.71 “All things being
equal,” he concluded, the appearance of greater need is “enough
to push us over the edge.”72
V. Recommendations for Reform
Reforms are possible at many levels. Obviously, the great
underlying problem is the national over-reliance on incarceration. The ultimate solution is to fix the criminal justice system.
Reforms are needed for sentencing, prevention, preparing people for return from prison, and a host of matters that other articles address. The nation needs to reconsider the relationship
between civil society and the hardware of justice, and truly to
address problems that are presently being locked up.
At the same time, the problems of the heartland cannot be
ignored. If they could, family farmers would prefer not to build
prisons on their land and towns would prefer not to exchange
the nighttime stars for the glare of perimeter lighting. But
without viable alternatives, such communities are pulling in
that direction. They will continue to pull until their fortunes
shift. Activists and scholars need to explore means to unite the
interests of urban progressive communities concerned about incarceration with rural progressive communities concerned
about the moral and economic health of the heartland.
The Census Bureau has the most power to make specific
reforms. In particular, people in prison should fill out their own
forms and provide what they consider to be their “usual residence.” This is the simplest, cleanest way to proceed. Most importantly, it will credit the genuine home community with
connection to these individuals.
It’s true that the institutional home may lose some formula
funding for water or sewerage, but these costs can be covered in
other ways. States often reimburse towns or counties for the
lost tax revenue associated with the public use of land, and
state departments of corrections often make arrangements for
71. Zachary R. Dowdy, Prisoner Count Tips Census Scales: Funds Don’t go to
Their Hometowns, NEWSDAY, Apr. 3, 2000, at A06.
72. Id.

\\server05\productn\P\PLR\24-2\PLR218.txt

606

unknown

PACE LAW REVIEW

Seq: 20

16-FEB-05

12:49

[Vol. 24:587

police and fire. These are the appropriate mechanisms for covering these costs.
It’s also true that the address provided might not be valid.
But court papers use an official “last known address” and the
parole authorities ask people in prison to provide an expected
address for return. If the address provided is good enough for
their purposes of individual tracking, it should be good enough
for the Census Bureau’s more statistical purposes. Moreover,
the Census Bureau accepts unstable addresses for homeless
people, migrant workers and even the highly mobile urban
youth who sometimes end up behind bars for a while. The difficulty in precisely identifying individuals’ home address should
not be a barrier to putting them in their own neighborhood
rather than an entirely different congressional district.
Funding formulas can also be adjusted to minimize this
problem. The Appalachian Regional Commission could adjust
the formula to “non-institutionalized population” rather than
“population.” The data is just as readily available from the Census Bureau, with just a few extra clicks on the web page. Similarly, agencies like HUD that assess minority representation or
per capita income can adjust their formulas to avoid the distortions of institutionalized population. It will enable them to target their limited funds more closely on the intended result.
Litigation might help to spur some particular reforms. The
Census Bureau’s use of the usual residence rule is likely not
“arbitrary and capricious” under the relevant legal authority.73
However, there may be some specific violations of equal protection. Stretches could be made to challenge funding flows, especially where the prison population is disproportionately
minority and people in prison are excluded by definition from
using the funds—such as specialty taxes for sport fishing.
Closer to fundamental rights, a plaintiff in a heavily minority
district with a large number of residents counted in prisons
outside the district may have a strong, if novel, vote dilution
claim. Even stronger arguments can be made under state con73. The Administrative Procedures Act gives the Census Bureau, like other
executive agencies, broad authority to determine its own rules. See 5 U.S.C. §706
(1966); see also Dist. of Columbia v. U.S. Dep’t of Commerce, 789 F. Supp. 1179
(1992).

\\server05\productn\P\PLR\24-2\PLR218.txt

2004]

unknown

Seq: 21

PRISONERS OF THE CENSUS

16-FEB-05

12:49

607

stitutional law, combining state equal protection theory with
state definitions of residence.
The U.S. Census Bureau consistently does a tremendous
job in the arduous task of the decennial count. It is odd, however, that people who are easy to count by virtue of their confinement should present such difficulty. A simple rule deprives
them of the economic and political clout to which they are entitled as members of this great nation. Fortunately, rules can
change.

\\server05\productn\P\PLR\24-2\PLR218.txt

unknown

Seq: 22

16-FEB-05

12:49