Prison-based gerrymandering in Tennessee Counties
The most dramatic instances of prison-based gerrymandering tend to be in rural counties that have large prisons, because a single state or federal prison can be the majority of a small county board district. The common solution used by more than 100 counties and municipalities is to remove the prison populations prior to redistricting. In most cases, this solution is permitted by law: Federal law gives local governments the right to determine the population base for their districts, and most state laws are silent on the question.
The operative word is “most.” Many of the most dramatic examples of prison-based gerrymandering are clustered in states where local governments have a little less flexibility when redistricting: notably Virginia, Wisconsin and the subject of this memo, Tennessee.
At the National Conference of State Legislatures in August, I met Tom Fleming, Director of the Office of Local Government within the Tennessee Comptroller’s office. He and I were attending the same session where the Census Bureau was soliciting feedback on the effectiveness of the Redistricting Data Program. After the session, he and I spoke about how the issue of prison-based gerrymandering effects Tennessee counties, and I offered to prepare a briefing memo on some possible solutions to the problem for this round of redistricting. That memo is below.
To: Tom Fleming
From: Peter Wagner and Aleks Kajstura, Prison Policy Initiative
Date: August 23, 2011
Re: Prison-based Gerrymandering in Tennessee Counties
This memo discusses the impact of the Census Bureau’s decision to count incarcerated people as residents of the prison location on Boards of County Commissioners districts in Tennessee, and discusses several ways that counties could harmonize the federal “one person one vote” principle with the unique requirements of the Tennessee State Constitution.
Introduction to the Problem
Every decade, counties in Tennessee, and across the nation, redistrict their Boards of County Commissioners to ensure equal representation in county government as required by U.S. Supreme Court precedent. When each district contains the same population, each member of the community is afforded equal representation. This undertaking, however, is vulnerable to any flaws in the data on which redistricting relies. There is a longstanding flaw in the Census that counts incarcerated people as residents of the prison location, even though almost all are barred from voting  and are not legal residents of the surrounding community. When district population counts rely on incarcerated populations, people who live close to the prison are given more of a say in government than everybody else. The practice of using prison populations to dilute the votes of other residents is referred to as “prison-based gerrymandering.”
One of the most dramatic examples in the nation that was created after the 2000 Census was in Lake County Tennessee where 88% of the population in County Commissioner District 1 was not local residents, but incarcerated people counted in the Northwest Correctional Complex. This gave every group of 3 residents in District 1 as much say in county affairs as 25 residents in other districts. Nine other counties in Tennessee saw dramatic instances of prison-based gerrymandering, as seen in this table, from the Prison Policy Initiative report Phantom Constituents in Tennessee’s Boards of County Commissioners:
|County||Most distorted district||Prison in district||Prison population (2000)||Vote enhancement|
|Bledsoe||District 1||Southeastern Tennessee State Regional Correctional Facility||969||34.70%|
|Davidson||District 20||Middle Tennessee Correctional Complex (now called the Charles Bass Correctional Complex), Riverbend Maximum Security Institution, and Lois M. DeBerry Special Needs Facility||2,569||15.40%|
|Hardeman||District 3||Whiteville Correctional Facility- CCA and Hardeman County Correctional Center||3,377||58.70%|
|Hickman||District 1||Turney Center Industrial Prison and Farm||1,106||33.40%|
|Johnson||District 5||Northeastern Correctional Complex||1,299||38.30%|
|Lake||District 1||Northwest Correctional Complex||1,799||87.60%|
|Lauderdale||District 5||Western Tennessee State Penitentiary||2,346||69.30%|
|Morgan||n/a||Brushy Mountain Correctional Complex||1,536||47.00%|
|Tipton||District 7||Tipton County Western Tennessee Detention Facility||527||9.50%|
|Wayne||District 2||South Central Correctional Facility and Wayne County Boot Camp||1,894||78.90%|
The Tennessee State Constitution may prohibit the most obvious solution
The U.S. Constitution requires counties to redistrict each decade, but in most states counties are free to choose which populations to include in the population base used to draw the districts. A few states require local governments to exclude prison populations when redistricting, and more than 100 counties and municipalities around the country independently choose to remove the prison populations when drawing local legislative districts. To assist counties in this process, the Census Bureau recently decided to produce the Advance Group Quarters Table to make the process easier.
Tennessee law, however, does not unambiguously give counties the same choices to define the population basis for their districts. In 1992, the Tennessee Attorney General concluded in an advisory opinion that the state constitution’s language that districts must “be based upon the most recent federal census” prohibits the removal of prison populations from the census data used for redistricting. The 1992 opinion reads, in relevant part:
It is the opinion of this Office that under the Tennessee Constitution… county legislative bodies must be reapportioned according to the 1990 Federal Census to include all persons enumerated in the Census.… Therefore, military personnel stationed at military installations in Tennessee, inmates at state correctional facilities and students attending colleges in Tennessee cannot be excluded from the population base in such reapportionment plans. (Tenn. Op. Atty. Gen. No. 92-21.)
This opinion, however, may no longer be reliable for two reasons:
First, had the Attorney General known that, 20 years later, Lake County would be forced to draw a district that is 88% incarcerated, giving the people who live next to the prison more than 8 times the influence of residents in other parts of the county, the Attorney General might have found a more flexible reading of “be based upon” that could have prevented such absurd results.
Second, the Census Bureau now provides a data product which is intended to help counties use federal census data to solve this exact problem. Prior to the 2010 Census, the only way to determine the population that the Census Bureau counted at a correctional facility was either by waiting until Summary File 1 was available, long after many states completed redistricting, or to just guess based on the location of the facility and information provided by the prison administration.
Twenty years ago, few people knew about the table in Summary File 1, but it was generally provided too late to use anyway. Most communities that excluded the prison populations during the 2000 round of redistricting used the second guesstimate method. Compared with the prospect of distorting their districts with a large population of incarcerated people incorrectly assigned to an individual district in their county, this method was superior; but under the Attorney General opinion this informal approach did not comport with the requirements of the Tennessee Constitution to use the federal Census. With the Advance Group Quarters Summary File, however, Tennessee counties are no longer faced with this problem of inadequate federal data.
Ideally, the Attorney General could revisit the 1992 opinion and come to a different conclusion today. And as we explain below, there are four ways that Tennessee counties can eliminate or minimize the vote dilutive impact of the Census Bureau’s prison counts without a change in state law.
Strategies to eliminate or greatly reduce the impact of prison-based gerrymandering in this round of redistricting
There are at least four ways that a Tennessee county can eliminate or reduce the vote dilutive impact of the Census Bureau’s prison counts. In order to use precise numbers in our explanations, we’ve chosen Morgan County as an example. These strategies are built upon the Prison Policy Initiative and Dēmos 2011 report Preventing Prison-Based Gerrymandering in Redistricting: What to Watch For.
Currently, in the districts drawn after the 2000 Census, Morgan County has a district where almost half of the “population” is located in the prison, giving the smaller number of actual residents in that district twice the influence of residents in other districts. We understand that the prison has grown over the last decade, and Morgan County is considering downsizing from an 18-person board (with six 3-member districts) to a 9-person board. Under the current 6-district system, the expanded prison population would be 64% of a district. If the prison population were used when drawing 9 single-member districts, Morgan County could be forced to draw one district that was entirely incarcerated. The possibility of this absurd result requires action.
Below, we review four options (one of which has two sub-options) and then explain how each is consistent with the requirements of the Tennessee and U.S. constitutions. The first two options completely eliminate the vote dilutive effects of the prison, and the second two options reduce it.
1. Use census data to identify and remove the prison population. A county can use the Census Bureau’s Advance Group Quarters Summary File to remove the prison population.
Analysis: The Tennessee Constitution requires that redistricting be “based upon the most recent federal census” but it does not mandate the choice of which federal Census product to use or how it should be used. A county can use the Census Bureau’s new Advance Group Quarters Summary File which was produced for the explicit purpose of assisting counties like Morgan.
Notably, while the Census Bureau counts incarcerated people as residents of the correctional facility, federal law does not require the use of any specific Census population set for redistricting and the U.S. Census Bureau does not mandate a single method of allocating incarcerated individuals for redistricting purposes, nor does it publish a single mandatory dataset. The Bureau provides various tools for redistricting, including a file produced for the first time this redistricting cycle for the purpose of empowering governments to make their own determinations as to the proper allocation of incarcerated and other populations for redistricting purposes. As the Census Bureau explained, the Advance Group Quarters Summary File will assist:
those in the redistricting community who must consider whether to include or exclude certain populations in redrawing boundaries as a result of state legislation. It will permit state and local redistricting officials to overlay this file with the 2010 Census Redistricting Data (Public Law 94-171) Summary File data.
As the Director of the Census Bureau further explained:
This decade we are releasing early counts of prisoners (and counts of other group quarters), so that states can leave the prisoners counted where the prisons are, delete them from the redistricting formulas, or assign them to some other locale.
Using federal census data to remove the prison populations prior to redistricting should be fully consistent with the Tennessee state constitutional requirement to base the districts “upon the most recent federal census.” However, in light of the existing Attorney General opinion, it would be best if the Tennessee Attorney General could be requested to review this issue in light of the Census Bureau’s new policy and provide a new opinion to this effect.
2. Draw the districts in such a way that the prison populations are included but do not distort the principle of “one person one vote.” Below, we describe two different methods to achieve the same end result, but some counties may find the differences between the methods significant:
2 A) Include the prison population, but draw the districts in such a way that the district with the prison is larger than the ideal size by the population of the prison. For example, Morgan County could draw 8 districts at about 2,182 and the ninth containing the prison at about 4,527. (Morgan County’s non-incarcerated population is 19,642, and divided by 9 districts is 2,182. The size of the ninth district is the ideal district size of 2,182 plus the 2,345 at the prison.)
Analysis: The State Constitution requires that the federal census be used, but does not mandate a maximum population deviation or specify the data to be used to evaluate that population deviation. Such a plan should be defensible under federal law because a rational basis exists for enacting a plan that had a large population deviation due to a non-resident, non-voting prison population.
State legislative districts are typically drawn to have a total deviation of less than 10% in order to rely on the presumption of constitutionality, but the deviation can be higher where there is a compelling state interest in doing so. Federal case law also recognizes the challenges of drawing county districts, and has approved plans with deviations over 10% at times. Although counties must comply with the “one person, one vote” principle, the courts have been sympathetic to the practical constraints that small counties face trying to draw equal districts.
The best way to satisfy the population equality requirement is to have a strong rationale for any population deviations. These need not be articulated in advance, but that is a best practice. It may be useful to cite the fact that over 100 local governments across the United States exclude prison populations while redistricting, and we consider Essex County New York’s language in their reapportionment law to be an excellent model:
“Persons incarcerated in state and federal correctional institutions live in a separate environment, do not participate in the life of Essex County and do not affect the social and economic character of the towns…. The inclusion of these federal and state correctional facility inmates unfairly dilutes the votes or voting weight of persons residing in other towns within Essex County….”
A plan drawn under this approach should be able to meet the state constitutional requirement to use the Census data for redistricting and comply with the federal “one person, one vote” requirement.
2 B) Include the prison populations, but not in any particular district. Treat them as “at-large” members of the county, similar to how state governments treat overseas military for state redistricting purposes. Draw 9 districts of 2,182, and then declare the prison population of 2,345 to be at-large. If someone in the prison registers to vote, the registrar can develop a procedure to assign them to a particular district.
Analysis: This is exactly how overseas military populations are counted for redistricting purposes. They are included in an “overseas” category for each state, but not assigned to individual blocks.
3. Split the prison into multiple pieces in order to lessen the impact of the vote enhancement caused by including the prison population in an individual district. Technically, Morgan County can physically split the block with the prison, or the county could reallocate some of the population to neighboring blocks that can be drawn into other districts. This latter idea is particularly appropriate for Morgan County, because the Bureau already appears to have counted the facility in the wrong location. The majority of the prison is not within the boundaries of Block 3007 Tract 110300, where the Census Bureau counted it, but is actually north and west of that location.) (See Figure 2.)
Analysis: This option will not eliminate the effect of the prison population, but it will reduce it. A two way split would result in two districts that are each 54% incarcerated. A three way split would result in three districts that are each 36% incarcerated. A four way split may be logistically more difficult in Morgan County, but it would result in four districts that are each 27% incarcerated. Numerous examples of split blocks exist across the country and we know of no federal or Tennessee prohibition against splitting blocks between districts.
Under this method, there is no practical difference whether the Block 3007 is split in to multiple pieces, or whether some of the population in that block is reallocated to neighboring Census blocks prior to redistricting. There may be technical reasons that the line drawing staff may prefer one method over the other, but they functionally lead to the same result.
4) Draw a multimember district that contains the prison so as to minimize the effect of the prison on equal representation. Draw 6 single member districts of about 2,182 residents and one 3 member district of about 6,546 that includes the prison. This district would be 36% incarcerated.
Analysis: In terms of the vote dilutive effect for Morgan County, this solution is roughly equivalent to the three-way split discussed above, and is superior to the two-way split considered above.
 Peter Wagner and JooHye DellaRocco, Phantom Constituents in Tennessee’s Boards of County Commissioners, Prison Policy Initiative (Feb. 21, 2008), available at: http://www.prisonersofthecensus.org/tncounties/.
 New York and Maryland have state laws that reallocate incarcerated people to their home addresses for redistricting purposes at the state, county and municipal levels. Colorado law requires all counties and school boards to exclude incarcerated populations when redistricting. Michigan law requires all counties and cities to exclude incarcerated populations when redistricting. New Jersey requires some school boards to exclude prisons, and the Mississippi Attorney General (Miss. Op. Atty. Gen No. 2000-0644 and Miss. Op. Atty. Gen No. 2002-0060) instructs counties to exclude the prison populations prior to redistricting.
 Prison Policy Initiative, Local governments that exclude prison populations, available at: http://www.prisonersofthecensus.org/local/
 The opinion also addressed the federal constitution and the Voting Rights Act. The Attorney General wrote that the “Federal Constitution may permit … [county legislative bodies to] exclude such groups from the population base in reapportioning those bodies.” The Attorney General said that exclusion “does not per se violate” the Voting Rights Act. (Tenn. Op. Atty. Gen. No. 92-21.).
 Available at: http://www.demos.org/pubs/Preventing_pbg.pdf.
 “Neither in Reynolds v. Sims nor in any other decision has this Court suggested that the States are required to include … persons denied the vote for conviction of crime in the apportionment base by which their legislators are distributed and against which compliance with the Equal Protection Clause is to be measured. The decision to include or exclude any such group involves choices about the nature of representation with which we have been shown no constitutionally founded reason to interfere.” Burns v. Richardson, 384 U.S. 73, 92 (1966).
 Available at http://www.census.gov/newsroom/releases/archives/2010_census/cb11-tps13.html.
 Robert Groves, Director, U.S. Census Bureau, So, How Do You Handle Prisons?, 2010 Census: The Director’s Blog (March 12, 2010). Available at http://directorsblog.blogs.census.gov/2010/03/01/so-how-do-you-handle-prisons/ .
 See generally National Conference of State Legislatures, Redistricting Law 2010 (2009). While courts presume that deviations under 10% are valid, district plans with smaller deviations have been struck down where plaintiffs showed that the districts resulted from unconstitutional actions. State of Tennessee Comptroller of the Treasury, Office of Local Government, A Guide to Local Redistricting in Tennessee, 2011 Edition, p.8-9.; see also Hulme v. Madison County, 188 F. Supp. 2d 1041 (S.D. Ill. 2011) (where no policy was offered to justify the deviation, and the court found that the districts were created to “satisfy the political agenda” of the chairman, to “cannibalize” his political opposition’s districts and included “threatening and coercive actions” against other Board members and so failed to “make ‘an honest and good faith effort to construct districts … as nearly of equal population as practicable…'” even a deviation below 10 percent was unconstitutional).
 “But as with most other cases in which governmental action is challenged as a violation of the federal Constitution, it should be enough that good reasons can be given for the action, whether or not they were articulated in advance of the action. E.g., FCC v. Beach Communications, Inc., 508 U.S. 307, 315, 113 S.Ct. 2096, 124 L.Ed.2d 211 (1993); McDonald v. Board of Election Commissioners, 394 U.S. 802, 809, 89 S.Ct. 1404, 22 L.Ed.2d 739 (1969)….” Frank v. Forest County, 336 F.3d 570, (7th Cir. 2003), cert. denied 540 U.S. 1106 (2004).
 Essex County Local Law No. 1 of 2003, available at http://www.prisonersofthecensus.org/nycounties/essex.html.
 Franklin v. Massachusetts, 505 U.S. 788, 806 (1992) (upholding the Census Bureau’s practice of counting oversees federal employees — mostly military — as “at-large” residents in their home states for Congressional reapportionment purposes even though the Census Bureau did not allocate them to individual blocks in that state).
 Alaska, District of Columbia, Indiana, Kentucky, New Hampshire, North Dakota, Pennsylvania, and Wisconsin all split blocks when drawing state legislative districts after the 2000 Census. For a list of these blocks, see http://www.census.gov/geo/www/sld/spblksld.txt.
 There is at least one case, however, that has struck down a legislative plan for relying on obvious Census errors. The Supreme Court of Oregon held that the Secretary of State was not obligated to rely on census data in apportioning districts. Hartung v. Bradbury, 33 P.3d 972, 598 (Or. 2001). Indeed, the court held that the Secretary of State violated the Oregon Constitution by failing to make corrections to federal census data to place a prison population in the correct census block. Id. at 599.