Prison gerrymandering dilutes the voting strength of urban areas where prisons are fewer and, thereby, weakens the political power of minority communities.

by Jyoti Jasrasaria, July 14, 2016

On July 19, 2015, the NAACP Legal Defense and Educational Fund, Inc. (LDF)—the nation’s oldest civil rights and racial justice law firm—submitted a comment letter in response to the Census Bureau’s Federal Register Notice about the Residence Rule and Residence Situations (Rule), 80 FR 28950 (May 20, 2015).

In its letter, LDF urges the Census Bureau to count incarcerated people as residents of their last known pre-arrest home address, rather than of the particular prison facility where they happen to be located on Census Day. According to LDF, first, this change would make the Rule more consistent with existing state laws, legal principles, and ordinary definitions regarding a person’s residence. Second, it would better align the Rule with the principles of an inclusive democracy by avoiding the inflation of political power in some communities (namely white rural communities where prison facilities tend to be located) over others (namely communities of color that are disproportionately represented in prison populations due to racial discrimination in the criminal system) when legislative districts are drawn. Third, an updated Rule would provide a more accurate picture of the nation.

For example, in 2010, 884 incarcerated Black individuals made up 56% of Martin County, Kentucky’s incarcerated population, but 12 Black residents made up only about 1% of the county’s non-incarcerated population. LDF asserts that these disparities in population are not only distressing, but also consequential for political participation:

[T]his artificial inflation dilutes the voting strength of urban areas where prisons are fewer and, thereby, weakens the political power of minority communities. This contravenes the constitutional principle of one person, one vote, which requires that everyone is represented equally in the political process, as well as the prohibition by the Voting Rights Act, now celebrating its [51st] anniversary year, on the dilution of the voting strength of minority communities.

LDF highlights the importance of an accurate Census count in light of its impact on redistricting. States and local jurisdictions often rely on Census data for redistricting purposes to carry out prison-based gerrymandering. This is the practice of counting incarcerated people where they are confined when drawing legislative districts rather than in their home communities, where they maintain enduring ties while they are imprisoned and where they likely will return. The average sentence for people incarcerated in state prisons is less than three years, but the redistricting and other ramifications that result from prison-based gerrymandering last an entire decade. Critically, counting incarcerated people as residents of their prison facilities is unjust given that incarcerated people are overwhelmingly denied the right to vote and do not interact with their surrounding communities by using parks and roads. LDF contends:

[P]rison-based gerrymandering is unlawful precisely because it artificially inflates population numbers, and thus, the political influence, of districts where prisons are located, at the expense of voters living in all other districts. Indeed, prison-based gerrymandering is all-too reminiscent of the infamous ‘three-fifths compromise,’ whereby enslaved and disfranchised African American people were counted to inflate the number of constituents in—and thus, the political influence of—Southern states before the Civil War.

As LDF’s letter explains, without an updated Rule, LDF and other organizations have been litigating and advocating for policy changes on a state-by-state and locality-by-locality basis to ensure compliance with the one-person, one-vote constitutional principle and the Voting Rights Act’s protection of minority communities’ voting strength. A Rule that counts everyone in the right place would help ensure a more robust democracy that gives all Americans equal representation in the political process.

Notwithstanding LDF’s urging, the Census Bureau’s proposed Rule has seemingly ignored LDF’s comments and maintained its practice of counting incarcerated people at their prison facility address.

Jyoti Jasrasaria, a 2L at Harvard Law School, is a 2016 summer legal intern at the NAACP Legal Defense & Educational Fund, Inc.


Incarcerated individuals often have little in common with the residents in the communities surrounding their correctional facilities, and elected officials do not always consider incarcerated people their constituents.

by Jyoti Jasrasaria, July 13, 2016

Justin Levitt sent a letter to the U.S. Census Bureau commenting on the Census Bureau’s Residence Rule and Residence Situations (Rule), 80 FR 28950 (May 20, 2015).

In his letter, Professor Levitt encouraged the Census Bureau to count incarcerated individuals at their last known address before incarceration as a means to further equal representation in the democratic process. As a professor of constitutional law and the law of democracy at Loyola Law School, as well as a practitioner and litigator in the area of political participation, Professor Levitt bases his comments on the structure of representation and the effects of various voting systems and districting plans.

Professor Levitt explains that the Census counts most people at their “home.” Those whose “usual residence” is different from their “home” are typically in a new location for work or education, “and they are generally intertwined with the communities where they are laying their heads most often” by interacting with their new neighbors, following community rules and regulations, and enjoying the benefits of local services and activities. However, this is not the case for the 2.2 million people in the United States who are incarcerated. Professor Levitt points to the fact that incarcerated individuals have little in common with the residents in the communities surrounding their correctional facilities. He notes the results of a recent study:

[T]here are now more than 450 counties where the proportion of African-Americans in the incarcerated population is larger than the proportion of African-Americans in the surrounding county — and more than 200 counties where the proportion of African-Americans in the incarcerated population is more than ten times larger than the proportion of African-Americans in the surrounding county.

Moreover, incarcerated individuals do not interact with the local community, and “most Village Township residents will not likely consider them ‘neighbors.’” Elected officials themselves do not always consider incarcerated people to be their constituents. As Professor Levitt recounts:

I]n 2002, a New York state legislator representing a district housing thousands of incarcerated individuals said that given a choice between the district’s cows and the district’s prisoners, he would “take his chances” with the cows, because “[t]hey would be more likely to vote for me.”

Indeed, according to Professor Levitt, 28 states have explicitly provided that incarcerated persons do not lose their residence in their home communities when they are incarcerated.

Professor Levitt asserts that the Census’s current practice creates (1) informational harm by promulgating misleading data about community demographics and (2) democratic harm by giving some communities disproportionate political influence and leaving others with diluted voting strength. In support of this point, Professor Levitt notes that 87% of the people living in one district of Lake County, Tennessee were incarcerated, which gave that district’s 344 non-incarcerated residents the same representation in county government as the 2500 or so individuals in the other two districts. Likewise, there are sometimes very few individuals eligible to run for office in districts with large incarcerated populations.

As Professor Levitt’s letter explains, many states and localities have attempted to compensate for the Census Bureau’s inadequate counting. The Census Bureau should assist these efforts by updating its current practice, which pre-dates the equal representation principle as well as the sharp increase in the incarcerated population. Moreover, according to Professor Levitt, there are feasible ways, which he suggests, such as surveying and questioning incarcerated people, just as the Census Bureau does for non-incarcerated people, to collect the information regarding an incarcerated individual’s last known address.

Despite Professor Levitt’s well-supported arguments for updated criteria for counting incarcerated populations, the Census Bureau has proposed to continue counting incarcerated people at their correctional facilities and, thus, to continue providing faulty demographic information that distorts political representation.

Jyoti Jasrasaria, a 2L at Harvard Law School, is a 2016 summer legal intern at the NAACP Legal Defense & Educational Fund, Inc.


The Census Bureau’s practice of counting incarcerated persons as “residents” of their correctional facilities, instead of their hometowns, harms all of Connecticut’s communities.

by Zack Goldberg, July 12, 2016

On July 17, 2015, Ms. Chandra Bozelko submitted a comment letter in response to the Census Bureau’s May 20, 2015 Federal Register Notice regarding the 2020 Decennial Census Residence Rule and Residence Situations.

Ms. Bozelko explained that she was formerly incarcerated at the York Correctional Institution in Niantic, Connecticut. Despite remaining registered to vote in her hometown of Orange, Connecticut, the Census Bureau’s residence rule counted her as a resident of the prison in Niantic.

Ms. Bozelko asserts that the Census Bureau’s practice of counting incarcerated persons as “residents” of their correctional facilities, instead of their hometowns, harms all of Connecticut’s communities – both those where prisons are located and the home communities that lose their residents to imprisonment. Ms. Bozelko explains: “counting [her] in one town when [she] was registered in another is a poor example of how we protect a citizen’s most treasured right.” Because “[t]he right to vote is sacred,” Ms. Bozelko urges the Census Bureau to count incarcerated persons as residents of their home addresses – and not as residents of their prisons’ addresses – in order to more accurately reflect the population of a given community.

Notwithstanding Ms. Bozelko’s experience and recommendation, the Census Bureau has proposed to maintain its policy of counting incarcerated individuals as residents of their prison facilities rather than their home communities.

To learn more about Ms. Bozelko’s leadership as a spokesperson for a Connecticut campaign to reform how incarcerated people are counted by the Census, click here. To read Ms. Bozelko’s recent op-ed on this issue, visit here.

Zack Goldberg, a 3L at Brooklyn Law School, is a 2016 summer intern at the NAACP Legal Defense & Educational Fund, Inc.


An example of prison gerrymandering in Wyoming shows that counting incarcerated people as residents of a prison produces inequitable results in the redistricting process.

by Zack Goldberg, July 11, 2016

On June 29, 2015, Professor Gerald Webster submitted a comment letter in response to the Census Bureau’s May 20, 2015 Federal Register Notice regarding the 2020 Decennial Census Residence Rule and Residence Situations.

As a professor, redistricting consultant for state and local governments, and expert witness in redistricting litigation, Professor Webster is a frequent user of census data. He asserts that counting the imprisoned population at the site of the prison can produce inequitable results in the redistricting process, commonly known as prison gerrymandering. Professor Webster contends that prison gerrymandering is particularly impactful in redistricting at the local level where jurisdictions with smaller total populations and significant imprisoned populations are given inflated voting power compared to those districts without prison populations. Additionally, Professor Webster contends that minority groups – who are disproportionately impacted by incarceration – are likely to have their voting strength diluted in the communities where the imprisoned populations come from as a result of prison gerrymandering.

Professor Webster provides an example of prison gerrymandering in his home state of Wyoming related to state senate districts 3 and 6. “To avoid having two incumbents in the same district, an appendage from district 6 is drawn north for 17 miles to include a prison housing approximately 500 individuals.” Since the ideal population for a Wyoming state senate district is approximately 19,000 individuals, the inclusion of 500 non-voting prisoners at the site of the prison artificially “inflates the value of ballots cast by non-prison voters in district 6 relative” to surrounding districts. A map of the prison gerrymandered districts, created by the Prison Policy Initiative, is available here. According to Professor Webster, this type of gerrymandering is unfair and undemocratic to the voters in the other 29 state senate districts.

Notwithstanding Professor Webster’s concerns, the Census Bureau has proposed to maintain its policy of counting incarcerated individuals as residents of their prison facility addresses rather than of their permanent home addresses.

Zack Goldberg, a 3L at Brooklyn Law School, is a 2016 summer intern at the NAACP Legal Defense & Educational Fund, Inc.


An expert on redistricting and population assessments makes the case for counting incarcerated people at their pre-arrest home addresses on Census Day.

by Harrison Stark, July 8, 2016

On July 18, 2015, Todd A. Breitbart submitted a letter commenting on the Census Bureau’s May 20, 2015 Federal Register Notice regarding the 2020 Decennial Census Residence Rule and Residence Situations. Mr. Breitbart, who has substantial experience in redistricting and population assessments in New York and California, contends that “[p]rison inmates should be counted as residents of their permanent home addresses, not at the places of incarceration.” Mr. Breitbart further compares the Census Bureau’s treatment of prisoners to other transient populations and argues that the Census Bureau is uniquely positioned to collect the data necessary to change their existing practice for counting imprisoned people.

Specifically, Mr. Breitbart compares the Census Bureau’s treatment of prisoners to the treatment of college or university students and individuals travelling for business or pleasure. Noting that students typically are counted at their temporary addresses (i.e., their schools), while travelers are counted at their permanent addresses (i.e., their homes they return to), he explains that the Bureau deviates from a strict “eat and sleep” interpretation so both groups can be counted at residences they voluntarily call home.

Moreover, Mr. Breitbart contends that there is a democratic deficit for the prisoners who cannot claim representation in the same way as students and travelers. As Mr. Breitbart points out, “members of Congress and state legislators, in furthering the interests of the permanent residents of their districts, also seek to further the interests of the students and visitors.” Indeed, representatives expand and support institutions of higher learning—for the benefit of students—and seek economic prosperity in their districts—to attract travelers. However, “[i]n contrast, no Congress member or state legislator seeks to represent the interests of the prisoners incarcerated in his or her district… To the degree that the prisoners enjoy representation in Congress or state legislatures, it is only from the representatives of the communities where they left behind their families and friends, to which they will eventually return, and where they may once again be voters.”

Moreover, Mr. Breitbart argues that urban communities like New York City, in particular, “[are] disadvantaged by the census rules relating to both prisoners and visitors.” As Mr. Breitbart shows, when New York reallocated its count of prisoners from “their places of incarceration” to “their permanent home addresses [in 2010], the population of New York City showed a net increase of 21,082,” a conservative number that did not take into account federal data. Mr. Breitbart furthered notes that “[New York C]ity’s population is permanently swollen by hundreds of thousands of visitors, but these persons are not counted here… if visitors are to be counted at their permanent home addresses, not where they are sleeping on Census Day, there is an even stronger argument for applying that principle to prisoners.”

Mr. Breitbart contends that the Census Bureau can change its practice of counting imprisoned people in the wrong place seamlessly and effectively. He explains:

[W]e now have extensive experience demonstrating that it would be quite practicable to count prisoners at their permanent home addresses. The states of New York and Maryland successfully adjusted their population databases for the 2010-12 redistricting without a huge investment of resources… These experiences can provide a model, and should lead the Census Bureau to reconsider its previous view that it would be prohibitively expensive to do what New York and Maryland accomplished.

According to Mr. Breitbart, the Census Bureau is uniquely situated to implement a change to how it counts imprisoned people. While individual states faced obstacles in collecting prisoner address information because of confidentiality concerns, Mr. Breitbart suggests that “the Census Bureau may well be able to address [the Federal Bureau of Prisons’] concerns about preserving confidentiality” and “will be in a far better position than the states, individually or collectively, to allocate prisoners to the census blocks of their permanent home addresses.”

Finally, Mr. Breitbart states: “Prisoners should be counted at the homes to which they will eventually return, where they left behind their families and friends, where they are represented by elected officials, and where they may once again be voters.”

Notwithstanding Mr. Breitbart’s recommendation, the Census Bureau has proposed to maintain its policy of counting incarcerated individuals as residents of their prison facility addresses rather than of their permanent home addresses.

Harrison Stark, a 2L at Yale Law School, is a 2016 summer intern at the NAACP Legal Defense & Educational Fund, Inc.


96% of the comments the Census Bureau received regarding residence rules for incarcerated people urged the Bureau to count them at their home addresses not at correctional facilities.

by Harrison Stark, July 7, 2016

On June 1, 2015, Jerry Zorsch submitted a comment letter in response to the Census Bureau’s May 20, 2015 Federal Register Notice regarding the 2020 Decennial Census Residence Rule and Residence Situations. Drawing on his experience as County Commissioner of Morgan County, Tennessee, including serving as a chairman of the County’s Redistricting Committee, Zorsch contends that because the Census Bureau “count[s] inmates in a prison in the county as if they were residents of that neighborhood,” counties like Morgan “end[] up with severe malapportionment.” The result is that “[i]t is impossible to count population bloc[s] like that in our county electoral system and still achieve equal representation among the citizens of this county.”

Commissioner Zorsch further explains that the Census Bureau’s residency designation contradicts Tennessee’s residency law. State law defines residence as “where the person’s habitation is fixed and is where, during periods of absence, the person definitely intends to return,” Mr. Zorsch contends, however, that imprisoned persons are not residents of their prison facilities in any meaningful ways:

These [incarcerated] men all come from outside our county. Upon release they immediately leave our county. They are not buying homes, raising families and putting down roots here. They came here, quite simply because they were forced to at gunpoint and they stay here only because of walls, wire and armed guards.

Moreover, Commissioner Zorsch demonstrates how the inclusion of prison populations by the Census Bureau has resulted in drastically unequal voting power for certain areas in Morgan County. Following the 2010 Census, the County’s correctional facility contained 2,400 inmates. Each district was drawn to contain 3,667 people, such that “whichever district [got] the prison block [would] only have 1,267 actual residents in it and 2,400 prisoners.” The result was that “the residents of one of our districts hav[e] 3 times the representation of the residents in the rest of the county.” Commissioner Zorsch explained how the problem got worse. One of two County prison facilities closed and the remaining facility absorbed the entire prison population, resulting in just one census block containing 10% of Morgan County’s population.

Finally, Commissioner Zorsch requests that the Census Bureau “please help us correct this problem and get back to the ‘One Man, One Vote’ ideal” and “help us to achieve fair and equal representation to all the citizens of our county, and those across this great nation by revising the Residence Rule or Residence Situations to count incarcerated people at home in the Census.”

Despite Commissioner Zorsch’s request, the Census Bureau has proposed to maintain the status quo and count incarcerated individuals at their prison facility addresses rather than at their pre-incarceration addresses.

Harrison Stark, a 2L at Yale Law School, is a 2016 summer intern at the NAACP Legal Defense & Educational Fund, Inc.


We offer our analysis.

by Peter Wagner, July 1, 2016

Yesterday, the Census Bureau disappointed hundreds of voting rights, civil rights and criminal justice reform organizations by declaring that it will continue to count incarcerated people at the facility, even though it causes prison gerrymandering.

The purpose of this blog post is to address the two positive proposals the Bureau made, put them into context and explain why they are nevertheless inadequate.

First, the context to yesterday’s disappointment. Last year the Census Bureau sought comment on its residence rules. In response, the Bureau was flooded with comments from urban and rural stakeholders explaining how prison gerrymandering harmed democracy in their state or local government and how the Census Bureau was in the best position to provide a remedy. In yesterday’s Federal Register Notice, which announced the new proposed rules the Bureau acknowledges that most of the comments they received were related to how the residence rule applies to people in prison, and that virtually all called for a change in the rule. They even summarized some of the arguments for change before declaring, without explanation, that they will continue to count incarcerated people in the wrong place.

The Bureau’s argument is that simply that “[s]tates are responsible for legislative redistricting” so it is not the Bureau’s fault that their flawed data undermines voting equality. Never mind that the data doesn’t meet the Bureau’s own goals for counting people “in the right place”, the Bureau will continue to collect data in a way that both the Bureau and many of their data users – and a growing number of federal courts — know is flawed.

There are, however, two encouraging announcements within the Federal Register notice that require some explanation and response:

  1. The Census Bureau re-affirmed its earlier commitment to publish group quarters data within the PL94-171 redistricting data. This information is essential to let redistricting officials know where prison facilities are — because this is more difficult than you might expect — and is critical to the efforts of some rural counties that wish to minimize the effects of prison gerrymandering by basing their districts on the population of their county not including the prison. For the 2010 Census, the Bureau agreed to publish this data much earlier than in previous decades, but they were not able to include it within the PL94-171 redistricting data. For this decade, the Bureau proposes to do exactly that, which will both make the data easier to find and will make the data more useful to jurisdictions like New Jersey and Louisiana that have extremely tight redistricting deadlines.
  2. The Bureau pledged to, on request, produce for states a special dataset that counts incarcerated people at home. The state would need to provide the necessary data, and the Bureau would produce a special dataset that the state could use. The Bureau doesn’t have any of the details on how this would work and the relevant requirements and schedules, but if this program worked smoothly, it could make it slightly easier for a state that wants to end prison gerrymandering to do so. (The experience of Delaware last decade both illustrates the value and pitfalls of this approach. The state had to postpone their law ending prison gerrymandering because the state claimed they couldn’t find a vendor to do the necessary adjustment in time. On the other hand, the state ran out of time because, like most states, they leave critical redistricting decisions to the last minute.)

Unfortunately, the Bureau’s proposal to shift responsibility for ending prison gerrymandering to the states is inadequate.

As we explained to the Census Bureau in our July 2015 comment letter:

The Census Bureau cannot leave fixing the prison miscount to the states.

However, all of this interest and activity [in hundreds of counties and dozens of states discussed at great detail earlier in the letter] in ending prison gerrymandering does not mean that the Census Bureau can leave this decision to the data users. As you know, the Massachusetts legislature concluded that that state’s constitution prohibits it from passing legislation ending prison gerrymandering. For that reason, the legislature sent you an earnest bipartisan resolution calling on you to count incarcerated people at home in the next census.[31]

These ad hoc solutions are even more out of reach for local governments. Many of the most dramatic instances of prison gerrymandering are concentrated in just a handful of states like Minnesota, Tennessee, and Wisconsin, where state constitutions or state law appear to prohibit the cities and counties from adjusting the Bureau’s data when drawing their districts without regard to the absurd and undemocratic results. For example, the Minnesota statutes declare “When used in reference to population, ‘population’ and ‘inhabitants’ mean that shown by the last preceding federal decennial census”[32] This results in cites like Waseca drawing city council districts that are 34.5% incarcerated, giving every 2 residents who live near the prison the political influence on city council of 3 residents in other parts of the city.

32. A copy of the resolution is at http://www.prisonersofthecensus.org/resolutions/MA-resolution-081414.pdf

33. Minn. Stat. Ann. &Sect; 645.44(8).

In addition, the Census Bureau’s proposal leaves out a key incarcerated group that state officials are unlikely to be able to collect home addresses for: people incarcerated in federal prisons. As Erika Wood explains in her report for Demos, Implementing Reform: How Maryland & New York Ended Prison Gerrymandering, neither Maryland or New York were able to reallocate people incarcerated in federal prisons. Federal Bureau of Prisons officials refused to share the necessary data with Maryland officials, and New York wrote its law on the correct assumption that federal prison officials would not cooperate with the state. State prisons are, without a doubt, the bulk of the correctional pie and the central nexus of the prison gerrymandering problem, but federal and military prisons in places like Kansas can be significant prison gerrymandering problems as well. The Bureau simply can’t shift responsibility for ending prison gerrymandering to the states and call the problem solved.


Talking points and submission instructions for comments to respond to the Census Bureau's proposal to count incarcerated people in the wrong place in 2020.

by Aleks Kajstura, July 1, 2016

UPDATE: Deadline extended to September 1.
The Census Bureau has extended the deadline to submit comments to September 1, 2016, explaining: “Because of the scope of the proposed criteria, and in response to individuals and organizations who have requested more time to review the proposed criteria, the Census Bureau has decided to extend the comment period for an additional 31 days.”

The U.S. Census Bureau released its proposed 2020 Census ‘residence rules’ that ignore overwhelming public input supporting a change in how incarcerated persons are counted in the Census. The Bureau announced it is leaving in place the inaccurate and outdated practice of counting incarcerated persons as "residents" of the prison location instead of their home communities. Interested stakeholders have until August 1 to submit comments before this proposal becomes final.

We’ve released our statement on these developments, and we’re working on our comment to the Bureau now, but it would be useful for the Bureau to hear from a diverse coalition of people. There are an impressive variety of reasons why the Bureau’s proposal to count incarcerated people at the location of the facility where they happen to be located on Census day is wrong and needs to be changed.

Here are a couple of talking points you can use to get started:

  • The Census Bureau is wrong to consider incarcerated people as residents of the correctional facility because [insert your own reason why people do not “live” in the prison.]
  • The Bureau has chosen to continue counting people in the wrong place, ensuring an inaccurate 2020 Census. [Describe how that impacts your community (counting incarcerated people as if they were “residents” of the correctional facility makes the Census less accurate for everyone: rural and urban communities; incarcerated persons and their families; governmental authorities trying to draw accurate redistricting plans; researchers trying to understand the demographics of local communities)]

Written comments regarding the proposed “2020 Census Residence Rule and Residence Situations” can be emailed by August 1 to Karen Humes, Chief, Population Division at POP.2020.Residence.Rule@census.gov

If you submit a comment letter to the Bureau, we’d greatly appreciate it if you could please forward a copy to us at FRN@prisonpolicy.org

You can also snail mail comments, but based on our experience with the Census Bureau and other federal agencies we strongly recommend using email. And if you need help putting together a comment letter with facts for your state, explore our website or contact us.



Stay Informed


Get the latest updates:



Share on 𝕏 Donate