The first African-American elected official in Somerset County Maryland explains how prison-based gerrymandering delayed electoral progress in his county.

by Peter Wagner, February 24, 2011

Last week Maryland’s WYPR Midday program with Dan Rodricks hosted a fascinating 49 minute-long discussion about Somerset County Maryland:

Craig Mathies is the first African-American ever elected to help run Somerset County, despite its one-third African-American population. A newly elected County Commissioner, Mathies joins Dan in Studio A to talk about his historic election and the challenges that lie ahead. Also joining us two other key players in the drama: Kirkland Hall, president of the Somerset County NAACP, and Deborah Jeon, legal director of the Maryland Chapter of the American Civil Liberties Union

Somerset County was a central part of the campaign against prison-based gerrymandering in Maryland because, as we explained in a fact sheet:

40% of the county is African-American, but an African-American has never been elected to county office. The county settled a Voting Rights Act lawsuit in the mid-1980s and agreed to draw a majority African-American district. Unfortunately, a new large prison in the remedial district resulted in the African-American resident population being split among multiple districts. An effective African-American district could be drawn if the prison population had not been included in the Somerset population count.

On the program, Craig Mathies, Kirkland Hall and Deborah Jeon explain the history of race relations and electoral fairness in the county and discuss their campaign for change.

You can listen online or download the audio.

A guide for redistricting advocates, shadow commissions and the media on avoiding prison-based gerrymandering.

by Brenda Wright and Peter Wagner, February 23, 2011

Prison-based gerrymandering is the practice of counting incarcerated persons as “residents” of a prison when drawing legislative districts in order to give extra influence to the districts that contain the prisons. The U.S. Constitution requires that election districts be roughly equal in size, so that everyone is represented equally in the political process. But prison-based gerrymandering distorts our democracy by artificially inflating the population numbers — and thus, the political clout — of districts with prisons, while diluting the political power of all other voters.

That this problem exists at all is largely an accident of two facts: (1) an outdated Census Bureau methodology that counts people in prison as residents of the correctional facilities, not of their legal home addresses; and (2) the skyrocketing rates of incarceration. Hopefully, in the future, the Census Bureau will eliminate the problem by counting incarcerated people as residents of their legal home addresses. Last year, three states — Maryland, Delaware and New York — had the foresight to pass legislation to eliminate prison-based gerrymandering within their borders. These three states now require that districts be based on Census data adjusted to reflect incarcerated people at their home addresses. More than a hundred rural counties and municipalities around the country have historically refused to engage in prison-based gerrymandering; they manually remove prison populations prior to drawing districts for local government. But most states and jurisdictions will still face the problem of prison-based gerrymandering in the upcoming round of redistricting.

When your legislature announces a proposed redistricting plan and invites public comment, you’ll need to act quickly to identify if and exactly how they used prisons to distort democracy in your state, county or city. This guide will tell you what to look for in the data and the state’s proposed plan in order to minimize the harm of prison-based gerrymandering.

(This guide assumes you have a mapping staff or sympathetic technical people on the redistricting body to assist you. Your technical allies can refer to our memo, Using the Census Bureau’s Advanced Group Quarters Table, which explains the timing, value, content and limitations of the Bureau’s prison count data.)

Protecting minority voting strength

Sometimes, a district that seems to have a majority-minority population really doesn’t, because of prison-based gerrymandering. If the minority “population” of the district consists of large number of incarcerated persons – who can’t vote – the district population numbers may be distorted. This creates districts that appear to give minorities the ability to elect the candidate of their choice, but in reality, they cannot. You need to examine any majority-minority district that includes a prison, to ensure that the district really has enough voting-eligible persons of color to create a viable majority with the ability to elect a candidate of choice to office.

Example: In order to settle a Voting Rights Act lawsuit, Somerset County Maryland intended to draw a district where African-Americans could elect a candidate of their choice after the 1990 and 2000 Censuses. But the inclusion of a large prison in the 1st Commission District split the sizable African-American resident voting population between two districts, leaving neither district able to elect a candidate of the African-American community’s choice. While the 1st Commission District appeared to be majority-African-American, in reality the district was not able to function as intended, because many of the purported African-American “residents” of the district were actually behind bars.

Similarly, although to different effect, prison populations sometimes create a false picture of racial and ethnic “diversity” within a district. Pointing out these examples is an effective way to raise the issue of prison-based gerrymandering and can be a powerful fact to raise if, as discussed in the next section, the state has under-populated districts that contain prisons.

Example: District 2B in Western Maryland drawn after the 2000 Census appears to be 15% African-American. But nearly all of that African-American population actually consists of incarcerated residents from other parts of the state who are unable to vote or to interact with the community in any way. The actual population of the district is overwhelmingly white.

Example: In 2002, the New York State Senate deliberately underpopulated districts in the upstate region while overpopulating districts in the downstate region. This problem ran parallel to the fact that the Census Bureau credited downstate residents to upstate census counts, and together served to dilute minority voting rights. For example, one of those upstate districts was the 59th Senate District, drawn to contain 294,256 people instead of the 306,072 that each district should have contained. Using Census data, the state reported that the district contained 6,273 African Americans, but three quarters of this population was incarcerated residents of other parts of the state. The legislature used the prison population to disguise the fact that the district had the smallest African-American population of any senate district in the state and they deliberately underpopulated that district to give it extra influence.

Keeping prison-based gerrymandering from making other malapportionment issues worse

Advocates should examine what percentage of each district is actually incarcerated, and how that interacts with the existing population deviations in the proposed districts. Keep in mind that in the strange world of redistricting, “underpopulated” districts have more political power than “overpopulated” districts, because in underpopulated districts, fewer people get the same opportunity to elect a representative as a larger number of people crowded into an “overpopulated” district. For that reason, a district that nominally falls within the 5% deviation rule applicable to state and local districts, but would fall outside that deviation without the prison population, should raise a red flag, and should be examined carefully to determine if the deviation should be reduced. Apart from that specific situation, any districts having large prisons should be scrutinized to avoid underpopulation of such districts compared to ideal district size, because including prison population magnifies the underpopulation of the district.

Note that the inverse is also a concern, even if we don’t have precise block-level data about the pre-incarceration home residence of people in prison who are currently being counted as “residents” of prisons. For example, you can use the fact that incarcerated people should have been counted at home to argue against extreme overpopulation of urban districts where incarcerated people disproportionately come from.

Limiting the vote enhancement in districts with prisons

Advocates should consider whether, if insufficient time remains to collect the home addresses of incarcerated people for this round of redistricting, the legislature can be persuaded to declare all incarcerated people to live at “unknown addresses” and not include them in the individual districts that contain prisons. (See Interview with Justin Levitt of the Brennan Center for Justice for more on why “unknown addresses” is superior to the Census Bureau’s status quo. )

If the legislature will not consider removing the prison populations from individual districts, advocates should examine ways to limit the magnitude of the vote enhancement to each district that contains a prison. Advocates should determine what percentage of each proposed district is actually incarcerated, and consider whether it is possible to configure the districts so that multiple large prisons are not concentrated in an individual district, thereby lessening the size of the vote enhancement in the prison districts. Similarly, if a single block contains a massive prison, advocates should consider whether the block could be split in two, so that the prison population can be placed in two different districts, thereby lessening the vote enhancement in any one district.


  • The Public Mapping Project is an open source redistricting package intended for shadow redistricting commissions and advocates. The software is building in support for advocates who wish to use alternative datasets, including Census data adjusted to remove prison populations, in their plans.
  • Using the Census Bureau’s Advanced Group Quarters Table
  • Data about prison-based gerrymandering, links to the database of historical (2005-2010) correctional statistics, and where, in May, we will post shapefiles with the Census Bureau’s group quarters data table and adjusted redistricting data that removes the prison populations and annotations of the blocks that contain correctional facilities with facility names, types, and more detailed demographic data.

Peter Wagner is Executive Director of the Prison Policy Initiative and Brenda Wright is director of the Democracy Program at Demos.

Actualización 28 de Marzo, 2012: Este articulo ahora está disponible en Español
Update March 28, 2012: This article is now available in Spanish.

Using prison populations to pad an Ark. Justice of the Peace district gives those residents 3 times the political power of residents in other districts.

by Peter Wagner, February 18, 2011

The Supreme Court said that districts must be drawn so that each person has the same access to government, regardless of where she lives. But a long standing flaw in the Census that counts incarcerated people in the wrong place undermines that principle.

One dramatic example is in Jackson County, Arkansas, where 64% of Justice of the Peace District 5 is incarcerated at the Grimes-McPherson Correctional Facility. Every group of 34 people who live in that district near the prison are given as much influence over the future of their county as 100 people in any of the other districts.

The solution? Ideally, the Census Bureau would count incarcerated people as residents of their home addresses; but Lincoln County, Lee County and Forrest City Arkansas all came upon a perfectly viable interim solution: don’t include the prison populations when drawing Justice of the Peace and city council districts.

Lincoln County didn’t really have much choice in the matter, as otherwise the prison would have been two districts with no voters all by itself.

Stay tuned for our research in other Arkansas counties. Interested in other states? See our list of local governments that exclude prison populations when drawing districts, and the list of prison-based gerrymandering problems we have identified.

New coalition says a fast redistricting timeline does not mean prison-based gerrymandering can not be addressed.

by Peter Wagner, February 16, 2011

New Jersey has the fastest redistricting timeline in the nation, but that doesn’t mean the state can’t address prison-based gerrymandering. That’s what members of the Integrated Justice Alliance have been telling legislators at hearings held across the state in recent weeks.

In their testimony, the Alliance called upon the state to follow the lead of Maryland, Delaware and New York and adjust the federal census to avoid prison-based gerrymandering. They’ve presented alternative ways for the state to avoid and minimize prison-based gerrymandering, and even presented the Apportionment Commission with the demographic data from the 12 Census blocks that contain the state’s federal and state prisons.

“Having presented testimony at four public hearings over the past two weeks, we have concrete indications that the public officials who sit on the Apportionment Commission of New Jersey have heard us and are ready to take some action during their redistricting process. It will be another victory for NJ in correcting racial and economic injustices,” said Alliance Chairwoman Margaret Atkins.

The Integrated Justice Alliance’s strategy around this and other criminal justice issues in NJ is being developed by a core steering group of 22 organizations who advocate for effective public policies before, during, and after incarceration in New Jersey. The Alliance grew out of another coalition that led an 18 month listening tour of community-based public hearings informed by those most directly affected by the criminal justice system around the state to craft and ultimately pass three major pieces of legislation addressing incarceration and reintegration policies and practices in New Jersey.

The Alliance’s submitted testimony, exhibits and contact information for the leaders are available on the New Jersey campaign page.

“I have found that most people don’t even realize that this kind of gerrymandering even takes place,” said Brother Aula Sumbry, Chair of the Racial and Economic Disparities Subcommittee of the Integrated Justice Alliance. “My prayer is that, as we make the citizens of this state and the Apportionment Commission aware of the fundamental unfairness of Prison Based Gerrymandering, the Commission will at minimum not continue the practice of padding the political power of the communities where prisons are located at the expense of the home communities of the prisoners and the rest of the state.”

Giving extra school board representation to a (former) prison town changes the outcome of a controversial vote in Maine.

by Peter Wagner, February 15, 2011

The decision last month by Maine’s Regional School Unit 13 to shift 8th graders in the town of St. George from the local school to a regional one is reinvigorating calls for an end to prison-based gerrymandering.

Next year, when 8th graders in the Maine town of St. George find themselves attending the 8th and 9th grade school in Thomaston instead of the local school, they’ll have prison-based gerrymandering to thank. On January 8, Maine’s Regional School Unit 13 decided by a very narrow vote — over the objections of the representatives from St. George — to move the 8th grade. The supporters of the school closure prevailed only because the representatives from Thomaston were able to cast additional votes because the town used to contain a prison.

As Brenda Wright, Aleks Kajstura and I explained in a letter to the Commissioner of the Maine Department of Education in 2009:

“[T]he Regional School Unit 13 apportionment is a system of weighted votes, where each town is given a number of votes in proportion to its population. However, the apportionment was conducted with Census data that did not reflect the actual population. The Regional School Unit based its apportionment on Census Bureau estimates for 2006 that credited the town of Thomaston with the population of the Maine State Prison that had closed 4 years prior.

As argued in Phantom Constituents in Maine’s Regional School Unit 13: How the Census Bureau’s Outdated Method of Counting Prisoners Harms Democracy, basing the weighted voting system on Census counts of prisoners at the now-closed Maine State Prison in Thomaston gives the actual residents of Thomaston an enhanced say over school board affairs.

By padding Thomaston’s actual resident population with the non-resident prison population, the current weighted voting system gives every group of 10 residents of Thomaston the same power over school district decisions as each group of 11 residents in the other towns.

With the RSU 13 board unwilling to request the Commissioner for permission to reapportion with corrected Census numbers, and the Commissioner refusing to make that determination on her own authority, board member Josiah Wilson has been collecting signatures as part of a petition campaign urging the Commissioner to order an new apportionment.

“The recent board decision to move the 8th graders is detrimental for my town and its school,” said board member Josiah Wilson. “But the fact that the prison counts influenced the outcome is making it easier to get voters all over the RSU 13 to sign the petition.”

Unlike the complicated redistricting process, the weighted voting formula required by Maine law is extremely straight forward. In the past, I helped Josiah Wilson calculate the votes for each town based on newer Census numbers that did not include the prison; and when the Census data for Maine are released, we’ll be able to calculate this within minutes.

The delay on ending prison-based gerrymandering in Maine is unfortunate, and I hope that via either Josiah Wilson’s petition or the pending appointment of a new Commissioner of Education, it ends soon. The prison has been gone for a long time. So too should prison-based gerrymandering.

A new prison in Adams County forces a choice in the Board of Supervisors.

by Peter Wagner, February 15, 2011

The Mississippi Natchez Democrat is reporting that:

NATCHEZ — The 2010 U.S. Census included a sizable chunk of people as Adams County residents who do not live in these parts by choice.

Approximately 2,000 inmates at the Corrections Corporation of America prison were counted as Adams County residents in the recent census, Adams County Correctional Center Warden Vance Laughlin confirmed….

Stacy Vidal, a public information officer for the U.S. Census Bureau, said each state is responsible for its own rules and processes for how they use census data to draw district lines….

She said some states elect to remove the prison population from its redistricting figures or use the data in other ways….

Board of Supervisors Attorney Bobby Cox said he thought the county had the option to exclude the prison population during the redistricting process.

He said he would have to consult with the Mississippi Attorney General’s office for guidelines on dealing with prison populations when redistricting. But he would think that since the inmates cannot vote, they should not be included, Cox said.

I recently wrote to most of the counties with large prisons in Mississippi to let them know about a 2002 Attorney General opinion instructing Wilkinson County that prison populations are not residents of the county and should not be included in county districts:

Inmates under the jurisdiction of the Mississippi Department of Corrections … are not deemed “residents” of that county or locality, as incarceration cannot be viewed as a voluntary abandonment of residency in one locale in favor of residency in the facility or jail. For purposes of the Census, these individuals should have been counted in their actual place of residence. Such inmates should not be used in determining the population of county supervisor districts for redistricting purposes by virtue of their temporary presence in a detention facility or jail in the county, unless their actual place of residence is also in the county. (Opinion No. 2002-0060; 2002 WL 321998 (Miss. A.G.))

The prison in Adams County was so new that we missed it in our initial preview of the 2010 Census, so the county did not receive this information by letter. I’ll follow up by phone with the County Attorney. The new prison is so large that if the county does not fix the Census, one district will be 30% incarcerated, giving every group of 7 people near the prison as much influence as 10 people in other parts of the county.

The Census Bureau's miscount complicates self-reflection in Vigo County and Terre Haute Indiana.

by Peter Wagner, February 13, 2011

My focus at the Prison Policy Initiative is on how the Census Bureau’s decision to count incarcerated people as residents of the prison’s location instead of as residents of their homes distorts the electoral process. But it’s also true that prison counts negatively impact planning in rural counties.

Six years ago, Rose Heyer and I wrote a report entitled Too big to ignore:
How counting people in prisons distorted Census 2000
that examined how the prison counts distorted county level statistics for total population, growth, race, ethnicity, gender and income. Our section on population growth appeared overly academic but may have been our most significant finding. Why? Because declining populations are often a sign of economic distress, and the increasing prison counts masked individual and regional problems. Our analysis of the 2000 Census found 56 counties that appeared to be growing but were actually losing population over the previous decade. Perhaps even more troubling, the misleading growth of West Texas counties with large prisons masked the consistent region-wide economic development problems faced in the region.

And that problem is repeating itself with the publication of this decade’s numbers:

At first glance, the Census Bureau statistics look pretty good. Vigo County’s population grew 1.9 percent to 107,848 in the 2010 count, up from 105,848 in 2000. Likewise, Terre Haute’s population jumped by 2 percent to 60,785 in the latest census, compared to 59,614 in 2000.

Initial stories about the Census in the Terre Haute paper were similarly celebratory of the population increase until I reached out to columnist Mark Bennett with whom I’d corresponded many times about census counts. This morning, he wrote:

But those increases contain an asterisk. Most of the growth can be attributed to the expansion of the Federal Correctional Complex and the additional inmates it now houses. (In the census, federal prison inmates count as residents of the community where they’re incarcerated.) In 2000, the local penitentiary held 1,764 prisoners at maximum- and minimum-security structures. A third facility was added in 2004. Thus, by Census Day 2010, Terre Haute’s federal inmate population had grown to 3,251, according to U.S. Bureau of Prisons statistics gathered by Peter Wagner of the Massachusetts-based Prison Policy Initiative.

His column, OUTTA HERE: Community effort needed to reverse trend in under-18 population calls for the county to establish an exodus czar to work with city and county officials and the “many energetic, progressive organizations” to “mount a full-scale attack on a trend threatening its future”. That trend is the number of families leaving the area.

The Census is an essential part of our democracy and the redistricting process, but’s it’s also an important tool to help communities see themselves and plan their future. Counting incarcerated people in the wrong spot just makes that difficult job harder. Congratulations to Mark Bennett for using the confusing numbers as a way to mobilize his community.

Sumter County Florida proposes to use prison populations to pad county legislative districts.

by Peter Wagner, February 8, 2011

The Orlando Sentinel is reporting that:

Sumter [County, Florida] officials say it’s likely they will use inmate-population figures from the 2010 Census to reapportion County Commission districts.

By my calculation, if Sumter County puts the newly expanded Coleman federal prison complex into the same district, they will be drawing a district that has more incarcerated people than local residents. This will give the voters in the district with the prison more than twice the influence of voters in other districts.

Over the past two months, I’ve been writing letters to county commissioners around the country, urging them to address prison-based gerrymandering. As part of that effort, we’ve also written fact sheets about prison-based gerrymandering in California counties, Florida counties, Georgia counties, Iowa cities and counties, Ohio cities, and Texas counties.

That consciousness-raising effort is bearing fruit, as can be seen in the final paragraphs of the same Orlando Sentinel article:

Bradford County in north central Florida — home to three state-prison facilities including Florida State Prison, where many death-row inmates are housed — also factors in close to 5,000 inmates in figuring its County Commission districts. According to 2009 U.S. Census estimates, Bradford’s population was 29,230.

Not everyone in that county agrees with counting prisoners.

Bradford County Commissioner Doyle Thomas said he plans to voice his objections when the panel discusses redrawing district lines in the coming months.

“I don’t think that’s right, because they [inmates] don’t vote and they don’t use services,” he said. “I’m going to say something about it.”

Letter to the editor asks why Senator Little wants to spend state funds to roll back the clock on prison-based gerrymandering.

by Peter Wagner, February 7, 2011

The Adirondack Daily Enterprise prints my letter to the editor Partisan prison gerrymandering:

To the editor:

I was disappointed to see that Senator Little wants to sue the state legislature to stop enactment of the prison-based gerrymandering law so that her district can claim prisoners as constituents. (“Little may challenge prisoner count law,” Feb. 1, 2011)

Part of me wonders if she is also thinking of suing her constituents in the Clinton, Essex, Franklin and Washington County governments for their past practice of refusing to use the prison populations when drawing local districts and weighting votes in the board of supervisors. Essex County even passed a local law in 2003 that said using the prison counts “unfairly dilutes the votes or voting weight of persons residing in other towns within Essex County.” Her constituents have been clear that prison-based gerrymandering is unacceptable.

But mostly I’m just sad to watch both parties in New York turn prison-based gerrymandering into a partisan issue. It doesn’t have to be this way. In Maryland and Delaware, similar legislation passed with bipartisan, urban and rural support. In fact, the lead sponsors in the Maryland Senate, Maryland House and Delaware Senate all had massive prisons in their districts. In those states, fairness trumped self-interest. Only in New York is spending state funds to roll back the clock considered a good idea.

Peter Wagner
Executive director
Prison Policy Initiative
Easthampton, Mass.

P.S.: The writer is the author of “Importing Constituents: Prisoners and Political Clout in New York.”

Proposed law would give more counties the choice to avoid using prison populations as padding.

by Peter Wagner, February 7, 2011

A bill that will give more counties in Virginia the option to avoid prison-based gerrymandering passed out of committee on Friday. It now goes to the full house.

Historically, Virginia law required counties to base their legislative districts on federal Census data, denying counties the flexibility exercised by counties in other states to choose the population basis of their required redistricting. In 2001, Virginia amended the law, giving counties where incarcerated people make up more than 12% of the Census population the option to draw districts without using the prison populations to pad the population of the district that contains the prisons. Of the 5 counties eligible under the law, 4 (Brunswick, Greensville, Richmond and Sussex) took advantage of it.

Unfortunately, the 2001 law did not give all counties relief. For example, the 2000 Census population of Southampton County was only 8% incarcerated, but when the county was divided in to 7 districts, the majority of one district was incarcerated. This gave the residents of the district with the prison more than twice the influence of the residents of other districts in the county.

Continue reading →

Stay Informed

Get the latest updates:

Share on 𝕏 Donate