The U.S. Census was originally used to determine the relative population of each state for Congressional apportionment, but another important use of the Census -- state and local redistricting -- is impaired by a very old quirk in how the U.S. Census counts people in prison. The Census Bureau counts prisoners as residents of the facilities in which they are incarcerated. The Bureau wants state and local government to use its findings, but it ignores the fact that the constitutions of many states, including New York, declare that a prison is not a residence. The U.S. Constitution and the Supreme Court's "one person one vote" cases requires state and county governments to redraw their districts each decade to give each voter the same access to government, regardless of where she or he lives.
There is a national controversy about how the Census Bureau's practice of counting thousands of disenfranchised non-residents in the rural Census blocks affects democracy at the state level, but a far larger result has gone almost unnoticed in the national press: the impact upon county governments. Because county legislative districts tend to be relatively small, a single large prison can have a huge negative impact on local democracy.
In this report, we analyze how the Census Bureau's method of counting prison populations affects voting strength in individual districts in the 31 New York counties with state prisons. Some counties have legislatures and draw equally-sized districts, while other counties have Boards of Supervisors and adjust the weight of the votes cast by individual Supervisors in proportion to the population they represent. For reasons of simplicity we will use the terms "legislature" and "redistricting" to refer to the various types of government and their redistricting and reapportionment processes.
We identified 13 counties that remove the prison population from their population base before redistricting and 16 counties where the legislature includes incarcerated people in the data used for redistricting. (See Table 1) In 5 of those counties, the impact of prisoners on equal representation is relatively small. In 5 other counties — Chautauqua, Livingston, Oneida, Madison, and St. Lawrence — including prisoners creates vote dilution in excess of 20%. Our county-by-county analysis can be found in Section IV, Summary of Findings in Individual Counties.
In two counties, Schuyler and Seneca, we were unable to classify the county as having excluded or included the prison populations. The prison in Schuyler County is relatively small and the existing population deviation between districts is relatively large. In Seneca County, the county has not updated its weighted voting system since the 1990 Census, which predates the two prisons in the county.
|Counties that include prisoners in population||Counties that exclude prisoners from population|
|Chautauqua County||Cayuga County|
|Chenango County||Chemung County|
|Columbia County||Clinton County|
|Erie County||Dutchess County|
|Fulton County||Essex County|
|Jefferson County||Franklin County|
|Livingston County||Greene County|
|Madison County||Orange County|
|Monroe County||Orleans County|
|Oneida County||Schoharie County|
|New York City||Sullivan County|
|Saratoga County||Washington County|
|St. Lawrence County||Wyoming County|
While counties must update their district lines each decade, New York State law gives counties the discretion to choose the population base used for redistricting. The decision to exclude the prison populations in the redistricting scheme therefore belongs to the County. While the majority of counties with large prisons choose to exclude the prison populations prior to redistricting, to our knowledge Essex County was the only county to publish a detailed justification for doing so. The Board of Supervisors wrote in Local Law No. 1 that prisoners "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" where they are incarcerated, and therefore cannot be residents. The law explains that counting prisoners as residents "unfairly dilutes the votes or voting weight" of Essex County's actual residents, and therefore the county will not include them in the population base. (The full text of the Essex County statute is included in this report.
The Census Bureau's decision to credit thousands of disenfranchised non-residents to the Census blocks with prisons creates serious problems for counties when they attempt to reapportion their local governments. New York counties with prisons are faced with a tough choice: adjust the federal census to ignore the prison populations, or rely on the Census and draw districts where some citizens are granted extra political clout because they happen to live next to a prison.
This problem need not exist in the future. Counties can lobby the Census Bureau for change or fix the data themselves; New York State legislators can also intervene.
In their recent report, the National Research Council of the National Academies recommended that the Census Bureau assess the states' interest in separately identifying the prison population in the PL94-171 Redistricting Data. Currently, correctional facility data is only available in Summary File 1, which is published three to five months after the PL94-171 redistricting data and in a different format. If the prison populations were identified in the PL94-171 redistricting data set, legislators could easily subtract the prisoners prior to redistricting and draw undistorted districts without needless difficulty or delay.
The National Research Council recommended that the Census Bureau assess states interest in identifying the prison populations in the redistricting data, but the impact of the prison miscount on counties is far larger than on states, so counties should make their needs known to the Census Bureau. The Census Bureau considers redistricting to be the second most important use of the Census, so the opinion of county officials in New York State would carry significant weight with the Bureau's Redistricting Data Program.
Counties that currently include the prison populations in their districts or weighted voting systems should redraw their districts or update their weighted voting systems. In Erie, Monroe, Westchester, Saratoga, and Ulster, the impact of including prisoners is relatively small, but the districts are still based on inaccurate representations of the counties' populations.
Including prisoners in the counties of Chautauqua, Livingston, Oneida, Madison, and St. Lawrence has significantly distorted local democracy. These counties should immediately redraw their legislative districts or re-weight their Supervisors' votes to ensure that all residents will have equal access to local government.
Restoring equal voting power in the counties of Chenango, Columbia, Fulton, Livingston, Madison, and Wayne, which have weighted voting, will be relatively simple, because only the votes for the Supervisors with prisons in their towns would need to be adjusted. (More detail on each county is given in Section IV, Summary of Findings.)
Counties could be spared expense and needless controversy if the New York Legislature passes the Prisoner Census Adjustment Act before the next Census in April 2010. The bill, written by New York State Senator Eric Schneiderman, directs the State Board of Elections to create a special version of the Census Bureau's PL94-171 Redistricting Data that counts prisoners where they are from, not where they are housed. The bill would mandate that the corrected dataset be used for congressional, state, and county redistricting.
Cayuga County excluded prisoners prior to drawing its 15 legislative districts; otherwise the district housing Auburn Correctional Facility would have been 32% prisoners.
Chautauqua County included prisoners when drawing districts for its 25-member legislature. District 23 (Portland) is 22% prisoners counted at the Lakeview Correctional Facility.
Chemung County excluded prisoners prior to drawing its 15 legislative districts; otherwise the district housing the Elmira Correctional Facility would have been 30% prisoners.
Chenango County included the prisoners counted at the Camp Pharsalia Correctional Facility when distributing votes to the Board of Supervisors. Each Supervisor represents a town and has a number of votes equal to about 1/21st the number of people counted in the Census. As a result of the prison population, the Plymouth Supervisor receives 97 votes instead of 85.
Clinton County excluded prisoners prior to drawing its 10 legislative districts; otherwise the district which includes Clinton Correctional Facility would have been more than 36% prisoners.
Columbia County included the prisoners counted at the Hudson Correctional Facility when distributing votes to the Board of Supervisors. Each Supervisor represents a town and has a number of votes equal to about 1/18th the number of people counted in the Census. As a result of the prison population, the Hudson Supervisor receives 432 votes instead of 386.
Dutchess County excluded the prison populations when it drew the districts for its 25-member legislature.
Erie County did not remove its prison population prior to drawing districts for its 15-member legislature. District 13 is 6% prisoners. That district, which contains Collins, Concord, Colden, Aurora, Orchard Park and the Cattaraugus Indian Reservation, includes the Collins and Gowanda Correctional Facilities in Collins.
Essex County excluded state and federal prisoners prior to distributing votes to the Board of Supervisors. Each Supervisor represents a town and has a number of votes equal to about 1/13th the number of residents — excluding prisoners — counted in the Census. Otherwise, the North Elba Supervisor would be able to exercise an extra 146 votes beyond her current 481.
Franklin County removed the prison population prior to drawing its 7 legislative districts; otherwise, the county would have had to create a district near Malone that was more prisoners than local residents. As three residents of the county explained in a 2004 letter to the Census Bureau: "Franklin County has always excluded state prisoners from the base figures used to draw our legislative districts. To do otherwise would contradict how we view our community and would lead to an absurd result...."
Fulton County included the prison population when distributing votes to the Board of Supervisors. Each Supervisor represents a town or city ward, and has a number of votes equal to 10% of the represented area's total population in the Census. The prison population counted at the Hale Creek Correctional Facility represents 6 of the 72 votes (8.5%) exercised by the Johnstown Supervisor.
Greene County removed the prison population prior to drawing the districts for its 14-member, 7-district legislature. Had the county included the 3,000 prisoners in the Coxsackie and Greene Correctional Facilities, the northeast corner of the county would have received 4 legislators rather than the 2 its resident population requires. Notably, the Legislature originally proposed including the prison population in the districts, but citizens elsewhere in the county strongly objected.
Jefferson County included prisoners when drawing districts for its 15-member legislature. The Cape Vincent Correctional Facility is 15% of the population in District 1. The prisoners counted at the Watertown Correctional Facility are 11% of District 11.
Livingston County included the prison population when distributing votes to the Board of Supervisors. Each Supervisor represents a town and has a number of votes equal to about 1/37th of the town's total population in the Census. Sixty-two percent of Groveland's Census population is incarcerated in the Livingston and Groveland Correctional Facilities. This enhancement of Groveland's population results in the Groveland Supervisor exercising 107 votes instead of 40. The Supervisor from Groveland is currently Chair of the Livingston County Board of Supervisors.
Madison County included the prison population when distributing votes to the Board of Supervisors. Each Supervisor represents a city ward or town and has a number of votes equal to about 1/45th of the represented area's total population in the Census. The prison population counted at Camp Georgetown represents 5 of the 21 votes exercised by the Georgetown Supervisor.
Monroe County is not seriously affected by where state prisoners are counted because the state prison in the county is very small (90 people) compared to an average district size of about 25,357 people.
Unlike upstate counties, New York City could not solve the problem created by miscounting prisoners solely by removing the prisoner populations from its districts. The Census does not give the City Council data on the many New York City residents incarcerated upstate, and so the Council is unable to put them back in the districts where they belong. The Census also counts the population of Rikers Island as part of District 22 in Astoria, Queens, where the prisoners are 8.1% of the population. Because Rikers Island is a local jail, not a state prison, its population belongs all over New York City, but the Council has no way of knowing which prisoners live where.
Oneida County included the prison population in the Census data used to draw districts for its 29-member legislature. The population of each district varies widely from the ideal district size of 8,119 people per legislative district. District 8 is about 17% too large, with a population of 9,469. District 28 is about 11% too small, with a population of 7,236. Both of these districts contain large prisons. The 3,604 prisoners counted at the Marcy and Mid-State Correctional Facilities in District 8 reduce the real population of the district to 5,865. Even more significantly, the smallest district, District 28, includes 2,451 prisoners counted at the Oneida and Mohawk Correctional Facilities. The resident population of District 28 is 4,785, 41% smaller than the average district in the county.
Orange County removed the state and federal prison populations prior to drawing its 21 legislative districts; otherwise, the state and federal prisons near Otisville would result in drawing a district that was more than 11% prisoners.
Although Orleans County claims to have included the prison population in its districts, the county in effect excluded the population of the Orleans and Albion Correctional Facilities from its districts. The County has a 7-member legislature. Three legislators must be from particular regions, but are elected at-large by the entire county and are therefore unaffected by where prisoners are counted. The remaining 4 legislators are elected from districts based on population. The County contains two state prisons, Albion and Orleans Correctional Facilities, both located in Albion. However, the Census Bureau incorrectly counted the Albion Correctional Facility in the town of Murray. To draw its districts, the county used corrected Census Bureau data so that both facilities were credited to Albion, where they are actually located. District 3 (Albion and Gaines) is 22% prisoners, but this does not enhance the weight of a vote in District 3 because the total population of District 3 is much larger than the other districts. In fact, the non-incarcerated population of District 3 is almost exactly what it should be if the correctional facilities were not included in the first place.
The St. Lawrence County legislature includes 3,120 state prisoners within only two of their 15 legislative districts. The population of District 2 is more than one-quarter prisoners. Although the legislative districts average 7,462 county residents, District 2 in Ogdensburg has only 5,639.
St. Lawrence County's inclusion of the prisoners was the subject of local controversy in part because the county had previously excluded the prisoners. The legislature changed course in order to shift the balance of power in the county. In response, about two thousand citizens signed a petition opposing their legislature's plan to include prisoners in population counts. At the county legislature's request, the county attorney scrutinized the petitions and succeeded in having enough signatures thrown out on technicalities to declare the petition invalid. The legislature went ahead with its districting plan, but many of the incumbents who were responsible for this were defeated in the following county election.
Saratoga County included the prison population when distributing votes to the Board of Supervisors. Each Supervisor receives the same number of votes as the Census Bureau's Count Question Resolution Program reported in the area represented by the Supervisor. The Mt. McGregor Correctional Facility is in the towns of Corinth, Moreau and Wilton. This prison population represents 247 of the 6,259 votes exercised by the Corinth Supervisor, 566 of the 13,549 votes exercised by the Moreau Supervisor and 27 of the 12,541 votes exercised by the Wilton Supervisor.
Seneca County has not updated its weighted voting system for the Board of Supervisors since the 1990 Census, when the prisons did not yet exist. The Willard Drug Treatment Center replaced the Willard State Hospital in 1995. The Five Points Correctional Facility in Romulus opened shortly after the Census was taken and, by July 2000, held more than 1,500 people. The County unsuccessfully lobbied the Census Bureau to include the Five Points Facility in the count, and debated hiring the Bureau to conduct a special Census. In the end, the county felt that the 2000 Census was not so different than the 1990 Census as to require updating the vote totals and retained its weighted voting system from 1990. While the County wanted the Census Bureau to include the Five Points Correctional Facility in its total population, the weighted voting system adopted after the 1990 Census excludes the temporary populations at the Willard State Hospital and the Seneca Army Depot in Varick.
After the 1990 Census, Seneca County excluded mental hospital and military populations prior to distributing votes to the Board of Supervisors. Each Supervisor represents a town and has a number of votes equal to about 1/44th the number of constituents — excluding hospital patients and military — counted in the Census. Otherwise, the Varick Supervisor would be able to exercise an additional 9 votes beyond her current 40.
The Schoharie County Board of Supervisors excluded the prison population at the Summit Correctional Facility when determining the number of votes each Supervisor would receive.
Schuyler County contains the Summit Shock Incarceration Correctional Facility in the town of Orange, in District 1. However, the way in which the districts were drawn makes it difficult to draw conclusions about how prisoners were counted. The County is divided into three multi-member districts, and there is a significant difference in the number of people per legislator, regardless of whether prisoners are excluded or included. This deviation, combined with the small size of the Summit Shock Correctional Facility, makes drawing conclusions difficult.
Note that the Census incorrectly classified the Summit Shock Incarceration Correctional Facility as a juvenile institution and therefore published its location in a different data tables (P37005) than the rest of the adult correctional center population in table P37003. We calculated the district populations as:
|District||Towns||Number of legislators||Total Population||Summit Shock Incarceration Population||Resident Population|
|1||Reading, Tyrone, Orange||2||5,252||301||4,951|
|3||Hector, Cayuta, Catherine||3||7,329||0||7,329|
Sullivan County excluded prisoners prior to drawing its 9 legislative districts; otherwise the district containing the Sullivan and Woodbourne Correctional Facilities would be 22% prisoners.
Ulster County included the prison population when drawing districts for its 12-district, 33-member legislature. Districts 1, 4 and 9 are 4.7 to 5.5% prisoners. Wallkill and Shawangunk Correctional Facilities are in District 9, and Eastern Correctional Facility is in District 1. The Ulster Correctional Facility is physically located near the Eastern Correctional Facility in Wawarsing, but the Census Bureau made a mistake and counted the facility in Saugerties. The county used the original Census Bureau data when drawing its districts, and distorted the weight of a vote in Saugerties' District 4 rather than adding to the distortion in Wawarsing's District 1.
Washington County excluded prisoners prior to distributing votes to the Board of Supervisors. Each Supervisor represents a town and has a number of votes equal to about 1/14th the number of residents -excluding prisoners — counted in the Census. Otherwise, the Fort Ann Supervisor would be able to exercise an additional 193 votes beyond her current 269.
Wayne County included the prisoners counted at Butler Correctional Facility in Butler and Walcott when distributing votes to the Board of Supervisors. Each Supervisor represents a town and has a number of votes equal to about 1/22nd the number of people counted in the Census. As a result of the prison populations, the Butler Supervisor receives 105 votes instead of 92. The Walcott supervisor receives 216 votes instead of 205.
Westchester County included the prison populations in Sing-Sing, Bedford Hills and Taconic Correctional Facilities when drawing its 17 legislative districts. The 2,279 state prisoners in Sing-Sing — 75% of whom are from New York City — are 4.2% of District 9's population.
Wyoming County excluded prisoners prior to distributing votes to the Board of Supervisors. Each Supervisor represents a town and has a number of votes equal to about 1/25th the number of residents — excluding prisoners — counted in the Census. Otherwise, the Attica Supervisor would be able to exercise an additional 87 votes beyond her current 164. Similarly, including the prisoners would have allowed the Perry Supervisor to exercise an additional 71 votes.
Previous research conducted by St. Lawrence County, the Greenville Press (Greene and So. Albany Counties), and Dan Jenkins has relied solely on interviews with county officials. This report used legal and demographic research to confirm how the existing districts were actually drawn, and used interviews to interpret and confirm our results.
In April, the Prison Policy Initiative published the Democracy Toolkit (available at http://www.prisonersof thecensus.org/toolkit/) to help rural counties determine whether prison populations were excluded from districts and to calculate the resulting vote dilution in counties that included them. We used this methodology and online tools in addition to the Arcview GIS mapping system.
Most counties did not publish the population totals for their districts, so we calculated the population totals by overlaying Census data over the county district maps in Arcview. In the counties that publish electronic files with the shapes of their districts, this calculation was straightforward. In other counties, we recreated their paper maps in Arcview and then calculated the population totals using both official Census data and a special version of the Census data with the prison populations removed. In this way, we could determine which dataset was used to draw the districts. Similarly, in counties with weighted voting, we calculated the number of total residents per vote and the number of non-incarcerated residents per vote for each town. In most cases, the numbers matched for one calculation and not the other. In the cases where neither set of numbers matched, we interviewed county officials to learn the source of the numbers used for their calculations.
This report was prepared with the support of The After Prison Initiative of the Open Society Institute. Arla Berman, Lindsey Frischer, Dan Jenkins and Allison Tompkins provided some initial research for this report. Andy Beveridge and Bill Cooper provided invaluable assistance locating and converting geographic data. Finally, many county clerks generously gave their time to help us understand how their local government was structured.
Peter Wagner is an attorney and executive director of the Prison Policy Initiative. In 2002, he authored the first district-by-district analysis of the impact of Census counts of prisoners on state legislative redistricting, Importing Constituents: Prisoners and Political Clout in New York (2002). He has presented his research at national and international conferences and meetings, including a Census Bureau Symposium, a meeting of the National Academies, and keynote addresses at Harvard and Brown Universities. His publications include, with Rose Heyer, Too Big to Ignore: How Counting People in Prisons Distorted Census 2000 (2004) and with Eric Lotke, Prisoners of the Census: Electoral and Financial Consequences of Counting Prisoners Where They Go, Not Where They Come From (2005).
Meghan Rudy is a law student at Hofstra University School of Law.
Ellie Happel graduated from New York University with a degree in Metropolitan Studies in May of 2005. Her senior thesis, Rikers Island and Representative Democracy: Collateral Consequences of Incarceration in New York City, won the Borgman Prize for the best senior thesis in the social sciences.
William Goldberg is a student at Smith College.
 "[N]o person shall be deemed to have gained or lost a residence, by reason of his presence or absence... while confined in any public prison." N.Y. Const. art. II, S 4.
 The maximum distortion, as measured by the size of the largest prison divided by the district population, in Erie County is 6%; in Ulster County 5.5%; in Westchester County 4.2%; in Saratoga County 4.1%; and in Monroe County 0.4%.
 New York Municipal Home Rule Law S10(1)(ii)(a)(13)(c): "As used in this subparagraph the term 'population' shall mean residents, citizens, or registered voters. A population base for such a plan of apportionment shall utilize the latest statistical information obtainable from an official enumeration done at the same time for all the residents, citizens, or registered voters of the local government. Such a plan may allocate, by extrapolation or any other rational method, such latest statistical information to representation areas or units of local government, provided that any plan containing such an allocation shall have annexed thereto as an appendix, a detailed explanation of the allocation."
 Essex County Local Law No. 1 of 2003.
 National Research Council, Once, Only Once, and in the Right Place: Residence Rules in the Decennial Census, September 2006, p. 250.
 The Census publishes counts of correctional facilities at the block level in Summary File 1, Table P37. Unfortunately, this table includes local jails, halfway houses and other types of facilities, but we offer online tools to determine which blocks are prisons in Section 2 of the Democracy Toolkit available at http://www.prisonersofthecensus.org/toolkit/section2.html. Advanced users of Census data may know that the Bureau classifies correctional facilities by type to the tract level in table PCT16, but in our experience we have frequently found these classifications inaccurate and recommend against relying solely upon PCT16 classifications. We instead recommend using our methodology which matches contemporary criminal justice data to the Census Bureau's results in order to classify facilities by type.
 Essex County Local Law No. 1 of 2003. See full text at http:/www.prisonersofthecensus.org/nycounties/essex.html.
 Letter from Daniel Jenkins, Mark Flack Wells, and Norman Gervais to the Director of the Census Bureau on July 9, 2004 in regards to the Establishment of the 2010 Census Redistricting Data Program, available at http://www.prisonersofthecensus.org/news/2004/09/06/ruralcitizens/.
 Greg Wright, "Census inmate count gives some areas more political clout," Gannett News Service, March 17, 2006.
 "Legislature remapping criticized," Daily Freeman, Jan 29, 2003 and letter at http://www.prisonpolicy.org/articles/freeman020303.html.
 Schoharie County Local Law #2-2002, March 15, 2002.