Prison Gerrymandering in Massachusetts: How the Census Bureau prison miscount invites phantom constituents to town meeting

By Aleks Kajstura
October 30, 2013

Executive Summary

When the first town meeting in the United States was held 380 years ago in Dorchester, Massachusetts, prison counts were probably the last thing on the participants’ minds. But today, the Census Bureau’s practice of tabulating incarcerated people as residents of prison locations, rather than of their homes, results in a subtle but significant distortion to the principle of “one person one vote.” This problem is called “prison gerrymandering” and it distorts local democracy in 7 Massachusetts towns that contain large correctional institutions and use the “representative town meeting” form of government.[1] The towns of Billerica, Dartmouth, Dedham, Framingham, Ludlow, Plymouth, and Walpole each contain a precinct where up to 35% of the precinct’s representatives are directly attributable to the Census Bureau’s prison miscount.

Federal, State, and local governments can take action to address prison gerrymandering. Ideally, the Census Bureau will solve the problem nationwide by agreeing to tabulate incarcerated people as residents of their home addresses in the decennial census. On the state level, the Massachusetts legislature should pass a pending resolution[2] that sends a strong message to the Census Bureau in support of a national solution to prison gerrymandering. In the meantime, this report proposes methods for town governments to minimize the effects of prison gerrymandering. Towns can redraw precinct lines to adjust the population totals for the prison miscount, or reapportion town meeting members between the precincts so that the number of members assigned to each precinct reflects actual constituent populations.

I. Introduction

The Supreme Court of the United States requires that voting districts contain equal numbers of people to ensure the constitutional guarantee of “one person, one vote”. Because populations shift over time, governments must periodically redraw their voting districts. Massachusetts cities and towns rely on the Census Bureau to provide accurate population data to use for reprecincting (redrawing precinct lines), but the data contains a major flaw: The Census Bureau counts incarcerated people at correctional facility locations rather than at their home addresses. When towns use this flawed data to apportion political power between wards or precincts, they unintentionally inflate the votes cast by people who live near large prisons while diluting the vote of every other resident. This is called “prison gerrymandering.”

The City of Gardner, for example, is divided into 5 wards. Each ward should have the same number of actual residents, but the incarcerated population that the Census Bureau counted at the North Central Correctional Institution located in Ward 1 accounts for a full quarter of that ward’s population. This means that any 3 people in Ward 1 have the same say in city government as 4 people in any other ward.

This problem affects both state and local governments in Massachusetts, but the greatest distortion appears in towns with a representative town meeting form of government. Prison gerrymandering’s harmful impact on local democracy in Massachusetts adds urgency to a resolution pending in the legislature calling on the Census to resolve the issue nationwide before the next redistricting cycle.

II. Prison gerrymandering in town government

Some of the most dramatic examples of prison gerrymandering in the state can be found in 7 Massachusetts towns. These 7 towns use a “representative town meeting” form of government and contain large correctional facilities.[3] In these 7 towns, I found that the Census Bureau’s method of counting incarcerated populations led to severe voting inequality:

  • In the Town of Billerica, the people incarcerated in the Middlesex County Jail & House of Correction are counted as if they were residents of Precinct 6. These incarcerated people account for 5 of the precinct’s 23 representatives at town meeting.
  • In the Town of Dartmouth, the people incarcerated in the Bristol County House of Correction and Jail, Bristol County Sheriff’s Office Women’s Center, and the C. Carlos Carreiro Immigration Detention Center are counted as if they were residents of Precinct 2. These incarcerated people account for 13 of the precinct’s 44 representatives at town meeting.
  • In the Town of Dedham, the people incarcerated in the Norfolk County Jail are counted as if they were residents of Precinct 1. These incarcerated people account for 7 of the precinct’s 39 representatives at town meeting.
  • In the Town of Framingham, the people incarcerated in the South Middlesex Correctional Center and MCI Framingham are counted as if they were residents of Precinct 16. These incarcerated people account for 2 of the precinct’s 12 representatives at town meeting.
  • In the Town of Ludlow, the people incarcerated at the Hampden County Correctional Center are counted as if they were residents of Precinct 5. These incarcerated people account for 5 of the precinct’s 15 representatives at town meeting.
  • In the Town of Plymouth, the people incarcerated at the Plymouth County Correctional Facility are counted as if they were residents of Precinct 10. The people incarcerated at the Plymouth County Correctional Facility account for 3 of the precinct’s 9 representatives at town meeting.[7]
  • In the Town of Walpole, the 482 people incarcerated at MCI Cedar Junction are counted as if they were residents of Precinct 5. These incarcerated people account for 3 of the precinct’s 18 representatives at town meeting.
Seven Massachusetts towns have town meeting members that represent phantom constituents
Precinct Representative Members Representatives attributable to the incarcerated population Percentage of a precinct’s representatives awarded due to the incarcerated population
Billerica, 6 23 5 21%
Dartmouth, 2 44 13 30%
Dedham, 1 39 7 17%
Framingham, 16 12 2 20%
Ludlow, 5 15 5 35%
Plymouth, 10 9 3 35%
Walpole, 5 18 3 17%

Despite the dramatic distortion caused by including the prison populations in reprecincting data, my interviews with state and local government officials revealed that the towns did not intentionally engage in prison gerrymandering. In fact, at least one town official was confident that incarcerated populations were not included in the precinct population totals while others just deferred to the Census data based on advice from the Secretary of the Commonwealth.

In Massachusetts, the towns’ reprecincting process involves the town’s Board of Selectmen, the Secretary of the Commonwealth’s office, and the LEDRC (Local Election District Review Commission). The towns often relied on the Secretary of the Commonwealth for technical mapping assistance and guidance on achieving population equality. The Secretary of the Commonwealth, however, operated with an eye toward the precincts’ role in forming state legislative districts, rather than the ensuring that the precincts fulfilled local needs.[8] This process resulted in precincts that include phantom constituents simply because the Census’ population data counts incarcerated people as residents of the location of the correctional facility, rather than their home address.

III. Solutions

Local governments, the state of Massachusetts, and the Census Bureau can all take steps toward solving the problem of prison gerrymandering.

What the Census Bureau can do

The obvious solution to a problem created by the Census Bureau’s methodology is to simply change the methodology. Counting incarcerated people at their home addresses in the decennial Census would solve the problem of prison gerrymandering across all 50 states. But state and local governments can also take action on their own. Currently four states have passed state-wide laws that reject the Census Bureau’s prison count methodology, requiring the states to adjust the Census Bureau’s data to tabulate incarcerated people in their home communities. At the local level, over 200 municipalities and counties do their own population adjustments to avoid prison gerrymandering when drawing their districts. A change in Census methodology would obviously make the redistricting process easier for those jurisdictions by avoiding the need to adjust the Census’ redistricting data. But even more importantly, a change at the Census Bureau would ensure that places such as these 7 towns in Massachusetts would never accidentally engage in prison gerrymandering again.

What Massachusetts is doing

The state of Massachusetts knows firsthand how important it is to find a solution to prison gerrymandering. After completing this decade’s state legislative redistricting, Massachusetts’ Special Joint Committee on Redistricting published a report analyzing the state’s redistricting experience. About a quarter of the report focused on how prison gerrymandering undermined the Committee’s efforts to draw fair districts. The Chairs of the Committee concluded that “the way prisoners are currently counted does a disservice to the state and should be changed.”[9]

To work towards a solution, the Chairs recommended that the legislature “adopt a resolution to be sent to the United States Congress and the Director of the U.S. Census Bureau expressing support for changing the residency classification for counting prisoners to their legal residence prior to incarceration.”[10]

And the Legislature is following the Committee’s recommendation: Senator Sonia Chang-Diaz and Representative (now Senator) Linda Dorcena Forry already introduced a resolution calling on the Census Bureau to count incarcerated people at home (bill S 309/H 3185). The resolution reaffirms the Redistricting Committee Chairs’ findings that “Census data results in distortions of the one-person, one-vote principle in drawing electoral districts in Massachusetts” (S309/H3185). At a March 27, 2013 hearing before the Legislature’s Joint Committee on Election Laws, a diverse group of voting rights and criminal justice organizations expressed strong support of the resolution.[11]

The resolution would add Massachusetts’ voice to the call for change issued by states that already reject the Census Bureau’s prison count methodology. It is especially important to states like Massachusetts to weigh in because the Census Bureau has a strong history of being responsive to states’ redistricting data needs, explicitly soliciting feedback every decade. The Census Bureau’s research and planning for the 2020 Census is already underway, so the Massachusetts legislature must pass a resolution in a timely manner to take advantage of this window of opportunity to shape the next census.[12]

Towns can adjust their populations for reprecincting purposes

More than 200 county, city, and town governments across the country[13] have independently avoided prison gerrymandering. Massachusetts towns have two options to prevent the Census Bureau’s method of counting incarcerated people from skewing local democracy:

Option 1: Redraw the town member precincts

  • Many local governments around the country have chosen to avoid prison gerrymandering by refusing to use the prison populations when drawing their local districts. Massachusetts towns can do the same.[14]

Option 2: Reapportion representative town meeting members

  • I note that at least three towns (Billerica, Dartmouth, and Walpole) can take advantage of a simple alternative solution: reassign some of the town meeting members to other precincts in order to compensate for the population distortions among precincts. For example, redistributing three members from the prison precinct in Walpole among precincts 6, 7, and 8 would eliminate the vote dilution caused by prison gerrymandering.

My research suggests that prison gerrymandering became a problem in Massachusetts local governments largely because the incarcerated populations included in the reprecincting data simply went undetected. Fortunately, the reprecincting process lends itself to a simple solution: guidance from the Secretary of the Commonwealth. As explained above, the Secretary of the Commonwealth plays a considerable role in town reprecincting. This provides a prime opportunity for the Secretary of the Commonwealth to provide information about prison gerrymandering and help towns to identify the prison populations in the reprecincting data so that towns don’t unintentionally undermine the principle of “one person, one vote.” I recommend that in the future the state should proactively help town governments avoid prison gerrymandering as part of the state’s role as a resource for towns in the reprecincting process.


The ideal solution to prison gerrymandering lies with the Census Bureau. A change in Census Bureau methodology to tabulate incarcerated people at their home addresses in the decennial census would provide a permanent nationwide solution to prison gerrymandering. The Massachusetts legislature needs to quickly pass its resolution calling on the Census Bureau to act before it is too late to solve the problem nationwide for the 2020 Census.[15] Meanwhile, Massachusetts towns can take action on their own when reapportioning town meeting members to ameliorate the effects of the Census Bureau’s prison miscount.


Thank you to our friend Brenda Wright at Dēmos for reviewing an early draft of the report. The research and publication of this report was supported by generous grants from the Open Society Foundations, the Public Welfare Foundation, and a network of individual donors who want to see an end to prison gerrymandering nationwide.


Appendix 1: Number of Representative Town Meeting Members, by town
Town Number of precincts Total number of members Members per precinct
Billerica 11 240 21-23
Dartmouth 9 390 43-44
Dedham 7 273 39
Framingham 18 216 12
Ludlow 6 112 14-15, with 23 at large
Plymouth 15 135 9
Walpole 8 150 18-19

Appendix 2: Population, by town
Town 2010 Census population Incarcerated population Facility
Billerica 40,243 798 Middlesex County Jail & House of Correction
Dartmouth 34,032 1,133 Bristol County House of Correction and Jail, Bristol County Sheriff’s Office Women’s Center, and the C. Carlos Carreiro Immigration Detention Center
Dedham 24,729 610 Norfolk County Jail
Framingham 68,318 763 So Middlesex Correctional Center, and MCI Framingham
Ludlow 21,103 1,228 Hampden County Correctional Center
Plymouth 56,468 1,516 Plymouth County Correctional Facility
Walpole 24,070 482[16] MCI Cedar Junction


[1] Today, towns with a Representative Town Meeting form of government redraw their town government precincts every decade in an attempt to comply with the basic democratic principle of "one person, one vote." But ever since the first Census was conducted in 1790, the Bureau has tabulated incarcerated people as if they were residents of the facilities where they are incarcerated, rather than at their home addresses. From 1790 through at least 1970, the Census Bureau's method of counting incarcerated populations did not impact redistricting simply because relatively few people were in prison. The unprecedented growth in incarcerated populations over the past few decades has magnified the way that Census data skews district populations, leading to increasingly dramatic vote dilution.

[2] S309 and H3185, "Resolutions Urging the Census Bureau to Provide Redistricting Data that Counts Prisoners in a Manner Consistent with the Principles of 'One Person, One Vote,'" sponsored by Senator Sonia Chang-Diaz and Representative Linda Dorcena Forry, respectively, introduced on January 22, 2013.

[3] While individual towns allocate representation in different ways, in all cases voting power is required to be distributed in accordance with the principles of "one person, one vote."

[4] I note that all of these towns have enough non-incarcerated actual population to cross the 6,000-person eligibility threshold, regardless that the Census bureau attributed the prison population to the town's total population.

[5] For more on representative town meetings, see the Secretary of the Commonwealth of Massachusetts' "Citizen's Guide to Town Meetings" at

[6] The way members are apportioned is not necessarily always the same from town to town. For example in Plymouth, every precinct elects 9 members, whereas precincts in Billerica range from 21 to 23 members, with the larger precincts having more members. See Appendix 1 for a breakdown of members in all 7 towns.

[7] Although the Census Bureau counted 197 people at MCI Plymouth, and that number was included in the total population for Precinct 5, the number of incarcerated people was very small compared to the actual population of the precinct. Thus the Census Bureau's prison counts at MCI Plymouth did not significantly distort representative apportionment.

[8] While the towns were ostensibly left to draw their precincts as they saw fit, the Secretary provided guidance on how to draw lines that would pass the LEDRC's mandatory review. So while the Secretary viewed its role as giving purely technical assistance without any policy implications, the towns were responsive to any suggestions from the Secretary and generally deferential to the Secretary's guidance.

[9] Senator Stanley Rosenberg and Representative Michael J. Moran, Report from the Chairs of the Special Joint Committee on Redistricting, p.17 (Dec. 12, 2012) available at [PDF]

[11] Leah Sakala, Mass. Committee on Election Law hears testimony in support of resolution to end prison gerrymandering, (March 28, 2013) available at

[12] The Redistricting Committee Chairs noted in their report that "the tabulation of prisoners should be at the forefront of Bureau priorities in evaluating and adjusting how the 2020 U.S. Census will be conducted." (Report from the Chairs, p.17 [PDF]) Although the next census is still over 6 years away, the Census Bureau needs to hear from Massachusetts that the state needs redistricting data that accurately reflects the residence of Massachusetts residents. The research forming the building blocks for the Census' fundamental methodologies has already began, but a narrow window of time for change still remains.

[13] A growing list of local governments that the Prison Policy Initiative has confirmed avoid prison gerrymandering is available at

[14] I note that all 7 of the towns have charters or bylaws that use the word "inhabitant" in describing precinct populations, which has led to confusion in Walpole about whether or not that language requires towns to engage in prison gerrymandering. (Officials debate counting prisoners in Walpole population, GateHouse News Service, Walpole Times, March 17, 2013, available at Our analysis doesn't support the concern that this language is a barrier to avoiding prison gerrymandering, however, because the Massachusetts Constitution itself sought to clarify the meaning of the word by tying the definition of "inhabitant" to a person's home: "And to remove all doubts concerning the meaning of the word 'inhabitant' in this constitution, every person shall be considered as an inhabitant, for the purpose of electing and being elected into any office, or place within this state, in that town, district or plantation where he dwelleth, or hath his home." (MA Const. Chapter I, Section II, Article II.)

Furthermore, the Supreme Judicial Court directly addressed the question of whether or not the federal census is consistent with the Massachusetts State constitutional definition of inhabitant in a 1974 Advisory opinion. Asked by the Massachusetts House of Representatives to determine whether the state census could use the "usual residence rule" of the federal census in conducting the state census, the Court issued a very clear "no." The Supreme Court confirmed that the federal census uses a definition of residence that ignores the Massachusetts Constitution's requirement that the term "inhabitant" implies place of domicile. ("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 her domicil." Opinion of the Justices, 365 Mass 661, 663-664 (1974).)

But regardless of whether a specific incarcerated person never intends to return home, state law bars that person from adopting the prison address as his or her residence. (I note the interesting continuing history of residence requirements for incarcerated people exercising their right to vote: In Dane v. Board of Registrars of Concord 374 Mass 152 (1977), the Supreme Judicial Court reiterated that incarcerated people are presumptively residents of their home districts and not of the district containing the prison, although the court did allow the Concord registrar to accept registrations from incarcerated people who had shown they had willingly established themselves as residents of the town of Concord. When people incarcerated in Norfolk failed to make such a showing, the Supreme Judicial Court upheld the refusal of the Norfolk registrar to register 619 people incarcerated at the Norfolk prison as residents of the town. Paul Ramos v. Board of Registrars of Voters of Norfolk, 374 Mass. 176 (1978) Subsequent legislation and Cepelonis v. Commonwealth, 389 Mass. 930 (1983) effectively removed this narrow discretion and required incarcerated individuals to vote via absentee ballot in their communities of origin.)

[15] The Massachusetts Redistricting Committee expressed concern that restrictive language in the Massachusetts Constitution (Article 119) makes it too difficult for Massachusetts to count incarcerated persons at their home address for state House and Senate redistricting purposes. The same restriction, however, does not apply to town governments. As such, towns are free to take action to avoid prison gerrymandering before the 2020 Census, as discussed above.

[16]MCI Cedar Junction straddles the Walpole/Norfolk town line. The Census Bureau counted 482 incarcerated people in block 2021, tract 411302, located in Walpole, and 246 incarcerated people in block 3048 tract 409101, across the town line in Norfolk. (The Census Bureau's method of counting incarcerated people as if they were residents of the location of the correctional facility does not skew democracy in Norfolk because the town does not elect town meeting representatives from precincts — Norfolk has an open town meeting form of government.)

Stay Informed

Get the latest updates:

Share on 𝕏 Donate