8
My first encounter with American gerrymandering came before I was even familiar with the term, or the concept for which it stood. I observed it every day when, as a young graduate student at the University at Buffalo in upstate New York, I climbed in my car to make the twenty-five-minute drive from my tiny duplex apartment—located in the town of Alden in the rural eastern part of Erie County—to the UB campus in Amherst, northeast of downtown.
Right along the street from where I lived were no fewer than three penitentiaries: the minimum-security Buffalo Correctional Facility, for some reason located fifteen miles outside the city of Buffalo; the Erie County Correctional Facility, whose website boasts of its “ ‘New Generation Jail’ pods and open bay construction,” whatever those are; and a maximum-security state prison, the Wende Correctional Facility, whose most famous inmates include Mark David Chapman, who murdered John Lennon in 1980, and Harvey Weinstein, the disgraced former media mogul. Most American prisons, especially those housing dangerous and violent criminals, are located away from major population centers. In Illinois, for example, 60 percent of inmates in the state correctional system are from Chicago, but 99 percent of them are serving their sentences elsewhere in the state.
In his 2017 book, Big House on the Prairie, the University of Wisconsin sociology professor John M. Eason estimates that 70 percent of the prisons constructed in the United States since 1970—a period that saw the number of correctional facilities more than triple—were located in rural communities. In that same time, the number of Americans incarcerated in prisons or jails also skyrocketed, from fewer than half a million to almost 2.5 million. When approximately 1 percent of your population is behind bars at any given time, the question of how to count those people for the purposes of political representation, and the distribution of federal funding and benefits, becomes a substantially more pressing one. Traditionally, both the U.S. Census Bureau and the governments of individual states have counted incarcerated persons not as residents of the counties, districts, cities, and towns where they lived prior to running afoul of the criminal justice system, and to which they will presumably return, but as residents of the facilities where they are being held.
The resulting overpopulation of rural areas, to which prisoners are disproportionately sent, and corresponding underpopulation of urban areas, from which they disproportionately originate, has profound implications for the division of political power. Exactly like the creeping gerrymanders of the early twentieth century, it creates a situation where citizens in rural areas have a much louder voice in the making of government policy than is justified by their numbers.
What I didn’t realize as I drove to campus those many years ago was that the more than two thousand inmates of the three correctional facilities I passed during my commute, despite not enjoying the right to vote themselves, represented 19 percent of the population of the town of Alden; 1 percent of the population of the 142nd State Assembly District; and together with those housed in other rural western New York facilities, 4 percent of the population of the Fifty-Ninth State Senate District. This phenomenon, whereby phantom constituents pad the populations of seats represented by rural politicians at the expense of their urban counterparts, is known as prison gerrymandering.
—
Danny Young won the 2005 election for the Ward 2 seat on the Anamosa City Council in Iowa by a single vote. This would be an unusual occurrence in any political race, even in a small town like Anamosa, which covers just 2.2 square miles and has only fifty-seven hundred residents. But it becomes even more remarkable when you consider that Danny, a fifty-three-year-old backhoe operator with no political experience, not only was never a candidate for the seat but raised no money, ran no campaign, and his name did not even appear on the ballot.
There were three reasons for his unexpected success. The first is that despite boasting a population of almost fourteen hundred people according to the 2000 census, Ward 2 contained only fifty-eight eligible voters, and most of them failed to turn out that November. The second is that the twelve-year incumbent, Bernie Keeney, had declined to run for reelection. Since no candidates had filed to replace him, this meant that write-in votes alone would determine who won. And the third is that only three ballots were cast in the Ward 2 contest: two for Danny, one of which was from his wife, the other from a neighbor; and one for Marty Seeley, his nominal opponent. Danny, now the city counselor elect, couldn’t even be bothered to vote for himself.
In the southeast corner of Ward 2 sits Anamosa State Penitentiary, the largest in Iowa. Its 1,321 inmates now represented 96 percent of Danny’s constituents. Or at least they did in theory. But with incarcerated persons being disenfranchised in all but two states, Maine and Vermont, there’s very little incentive for the politicians who represent districts whose populations are artificially inflated by the inclusion of prisoners to pay any heed to their liberty-challenged inhabitants.
Map of Anamosa, Iowa, showing the location of Anamosa State Penitentiary.
“Do I consider them my constituents?” Danny pondered in an interview with The New York Times, before amply summarizing the representational deficiencies that prison gerrymandering creates: “They don’t vote, so, I guess, not really.” The man he replaced as Ward 2’s representative on the city council gave a more polished politician’s answer when asked the same question. “A lot of the things we do in the community does affect them,” Bernie Keeney told The Anamosa Gazette. “We have to serve all citizens.”
Anamosa is by no means an isolated example. In almost every state, prisoners are routinely counted as residents of the institutions where they are incarcerated. The resulting representational discrepancies have been extensively documented by the Prison Gerrymandering Project, a nonpartisan effort launched by the Prison Policy Initiative think tank in the early years of the twenty-first century to raise awareness of the problem and lobby for reform.
Beginning in 2002, the organization produced a series of state-by-state reports highlighting some of the most dramatic effects of the prison-industrial complex on America’s legislative elections. What they discovered is that prison gerrymandering, which had largely flown under the radar as the War on Drugs precipitated the escalating mass incarceration crisis—one that every year sees 600,000 individuals pass through the gates of the nation’s two thousand federal and state prisons, and 10.6 million occupying cells in its three thousand local jails—was having very real effects on the allocation of resources and political power.
In New York, for example, the 2000 census saw some forty-four thousand residents of New York City counted in rural upstate areas. These included the more than two thousand inmates who resided down the street from my apartment in Alden. The imbalance was sufficient to deny the city an additional state senate seat, which the Republican majority was instead able to relocate to an upstate area where they could be reasonably assured of controlling it. Two of the politicians whose districts boasted the largest populations of prisoners, “representing” fully 23 percent of the total incarcerated persons in the state, were the Republican senators Dale Volker and Michael Nozzolio. These men also happened to chair the two committees in charge of criminal justice policy. Whenever momentum began to build in Albany to repeal the state’s harsh drug laws—laws that sent a constant parade of nonviolent offenders, and taxpayer dollars, to their districts—Volker and Nozzolio managed to short-circuit it.
The most extreme example, however, was New York’s 114th State Assembly District. Located in the far northern reaches of the state along the Canadian border, it is home to two large maximum-security prisons: the Upstate Correctional Facility and the Clinton Correctional Facility. Fully 7 percent of the district’s inhabitants at the time of the 2000 census—including 83 percent of its African American adult residents and 74 percent of its Hispanic residents—were disenfranchised prisoners.
By decade’s end, the prison population of the 114th would also include its state assemblyman, the Republican Chris Ortloff. In 2008, the Vietnam veteran who had represented the seat since 1986 was arrested by federal authorities at a motel in Colonie, New York, on felony charges of soliciting sex with a minor. The indictment alleged that he had made arrangements to meet two sisters, aged eleven and twelve, whom he believed he had been grooming online. In reality, he was chatting with an undercover state police officer. At least he was spared the indignity of being incarcerated in his own district. After pleading guilty, Ortloff was sent to the federal correctional facility in Danbury, Connecticut, to serve a minimum ten-year sentence. He was also forced, somewhat ironically, to resign his $110,000-a-year positions on the state parole board and the legislature’s redistricting taskforce. As will be discussed in chapter 10, he was far from the only member of the New York legislature to find themselves entangled in the criminal justice system.
Then take the case of Michael Cady, a longtime vagrant who in the fall of 1893 was convicted by the state of illegally registering to vote. Cady had an unusual habit. Every six months or so, he would visit the local police constabulary and confess to the crime of vagrancy. Under New York law at the time, this offense did not lead to a loss of voting rights. But it did allow a magistrate judge to commit the offender to a correctional facility for a set period of time. Cady would then be sent to the Tombs—the colloquial name for what is now the Manhattan Detention Complex, and what was then the New York City Halls of Justice and House of Detention—the notorious city jail in the Five Points neighborhood of lower Manhattan. There he would be provided with room and board, and the warden would employ him to run errands, sometimes even outside the prison walls. Cady had been doing this for seven years and apparently had no intention of stopping anytime soon.
Despite his unusual lifestyle, Cady evidently still felt a sense of civic responsibility, for he registered to vote at what he at least considered his primary residence: the Tombs. But the authorities saw things differently. Even though the federal census, which the state used to draw its legislative districts, counted him as a resident at the location where he was incarcerated, the state constitution, which provided that “no person shall be deemed to have gained or lost a residence, by reason of his presence or absence…while confined in any public prison,” did not. Cady was convicted of illegal registration, and the New York Court of Appeals unanimously affirmed. The court documents unfortunately make no reference as to whether he was permitted to serve his sentence at the Tombs.
Outside New York, a quick perusal of the results of the 2000 census—subsequently used by all fifty states to redraw their legislative districts—reveals some additional disturbing abnormalities. There were twenty-one U.S. counties, eight in Texas alone, where more than one-fifth of the reported residents were incarcerated prisoners. Those behind bars made up 33 percent of the population of Concho County, Texas (which, in addition to having the fourth-lowest per capita income in the nation, is home to what was then Eden Detention Center, since closed and repurposed as an Immigration and Customs Enforcement facility); 33 percent of the population of West Feliciana Parish, Louisiana (location of the Louisiana State Penitentiary, the largest maximum-security prison in the United States, also known as the Alcatraz of the South); and 35 percent of the population of Crowley County, Colorado (site of the privately owned and operated Crowley County Correctional Facility, which in 2013 paid $600,000 to settle a class-action lawsuit filed by a group of prisoners who were injured in a 2004 riot that the chronically understaffed facility had failed to adequately respond to).
Number one on this list would be Union County, Florida. Thirty percent of its 5,500 residents were inmates at the Union Correctional Institution, part of the sprawling Florida State Prison complex where the state’s death row is located and Ted Bundy was executed by lethal injection in 1989. But the property is bifurcated by the New River, placing about half of the inmates across the border in Bradford County, where they also find themselves in an entirely different congressional district. Fifty-six counties whose populations declined between 1990 and 2000 appeared to be growing according to the census numbers, solely by virtue of their increasing prison populations. Jones County, Texas, which lost 355 actual residents, saw its census population rise by 26 percent, thanks to the 4,650 additional inmates sent their way by the Texas Department of Corrections.
Unlike the gerrymanders discussed in chapter 6, the phenomenon of prison gerrymandering was not a product of overt racism. Nor was there any deliberate attempt among policy makers to distort the populations of legislative districts, at least initially. Instead, it came about largely as an accident of history. When the very first federal census was conducted in 1790, counting prisoners at their place of incarceration was not unreasonable. “While incarceration rates were low, prisons were in or near prisoners’ home communities, and census data was used solely to allocate congressional seats, prison gerrymanders generate[d] little controversy,” writes the DePaul University political science professor Christina Rivers. “They also had minimal influence on representational outcomes.”
But the policy of counting prisoners at their “usual residence,” defined by the Census Bureau as “the place where a person lives and sleeps most of the time,” has become considerably more problematic in the era of mass incarceration. “Current census residency rules ignore the reality of prison life,” argues Kenneth Prewitt, a former director of the Census Bureau under President Bill Clinton. “Incarcerated people have virtually no contact with the community surrounding the prison. Upon release the vast majority return to the community in which they lived prior to incarceration.”
Nevertheless, while fully 62 percent of the public comments received in response to the proposed residency rules for the 2010 census related to the counting of prisoners, the Obama administration declined to adjust its formula. In the lead-up to the 2020 census—citing the “major operational issues for both the correctional facilities and the Census Bureau” that would result from a change in policy—the Trump administration followed suit. One wonders whether those “major operational issues” might be solved, to throw an idea out there, by simply asking those who administer the nation’s jails, prisons, and detention centers for the last known address on file for each of their inmates. The handful of states that on their own initiative have opted to adjust the census data to count prisoners at their homes, rather than their cells, did not appear to have any significant difficulty overcoming the logistical challenge.
The negative antidemocratic effects of prison gerrymandering cannot be ignored. Not only does it create urban-rural imbalances in the comparative level of representation that different communities receive in state and federal government, but these in turn affect the allocation of federal and state tax dollars and other resources. In Maryland, for example, a report by the Prison Gerrymandering Project documented that “18% of the population currently credited to House of Delegates District 2B (near Hagerstown) is actually incarcerated people from other parts of the state. In effect, by using uncorrected Census data to draw legislative districts, the legislature granted every group of 82 residents in this district as much political influence as 100 residents of every other district.” The result: both power and money flowed away from the city of Baltimore and into the sparsely populated Appalachian counties of western Maryland.
Things were even worse in Somerset County, located along the Eastern Shore of the Chesapeake Bay. “A large prison is 64% of the 1st County Commission District,” the report continues, “giving each resident in that district 2.7 times as much influence as residents in other districts. Even more troubling is that by including the prison population as ‘residents’ in county districts, the county has been unable to draw an effective majority–African American district and has had no African-American elected to county government, despite settlement of a vote dilution lawsuit in the 1980s.”
That lawsuit had ended the practice of at-large districting, which had prevented the county’s African American residents—who make up 42 percent of its population—from being able to elect even a single black commissioner. But soon after the lawsuit had settled, the brand-new Eastern Correctional Institution opened its doors. The resulting impact on the 1990 census numbers turned what appeared to be a majority-minority district into one that, once disenfranchised felons were excluded, actually had a sizable white majority.
“Another problem with prison gerrymanders,” Rivers writes, “is that of ‘ghost’ constituents. Electoral district populations are not adjusted with the release or relocation of incarcerated individuals. Because newly released individuals are not counted at their new address until the next census, their political presence remains incarcerated at their former prison address. The same would go for those who are transferred from one institution to another.” So whatever inequalities are created by the counting of imprisoned persons at the time of the census are frozen in place for the remainder of the decade. An entire prison may be closed down, and every inmate relocated out of both the county and the districts of which it is part, while the remaining residents continue to reap the benefits of its inflationary effect on their numbers. “Consequently, though no longer physically present in that district,” she continues, “these ‘phantom’ constituents continue to enhance the political power of the communities that count them in their districts for many years.”
All of this would be disturbing enough even if the demographics of the incarcerated population perfectly matched those of the public at large. At least then the negative effects of prison gerrymandering would be equitably distributed across different demographic subgroups. But the problem is far worse than that. As the Somerset County example illustrates, the effects of prison gerrymandering are felt most acutely not by the nation’s white citizens but by its more vulnerable and marginalized minority populations. The downstream effects of the shocking racial inequities that exist within the American criminal justice system—themselves a product of systemic racism and a lack of access to educational and other resources that contribute to the lower socioeconomic status of many minority communities—have been a massive overrepresentation of African Americans and Latinos in the U.S. prison population. And with prisoner disenfranchisement being the norm, the sum effect of prison gerrymandering has been a decades-long transfer of political power away from America’s African American and Latino residents and into the hands of the white majority.
Take rural Brown County, Illinois, located west of Springfield between the Illinois and the Mississippi Rivers. According to the 2000 census, Brown County’s population was 18 percent black. But all but five of those 1,265 African American residents, 99.6 percent of the total, were inmates at the Western Illinois Correctional Center. The facility’s 2,000 prisoners made up almost a third of the county’s population. In the nation as a whole, there were 173 counties where more than half of the black population was incarcerated at the time of the 2000 census.
As mentioned earlier, the origins of prison gerrymandering were largely a historical accident, a product of decisions made during the very earliest years of the Republic that became entrenched due to bureaucratic inertia and governmental convention. And while that remains true, in more recent decades politicians have also embraced prison gerrymandering as a deliberate strategy. In a 2012 study, the Princeton University political scientist Jason P. Kelly analyzed what happened to prison populations in states where control of redistricting had flipped from one political party to the other. Recognizing that these situations provided the most fertile opportunities for politicians to manipulate prison populations for political gain, he hypothesized that “by shifting a significant proportion of these phantom constituents into districts that lean heavily toward the majority party, legislators can free up an equal number of citizens from those districts to be distributed among neighboring marginal ones, thereby increasing that party’s likelihood of picking up additional seats in the state legislature.”
“Alternatively,” he continues, “prison populations in relatively safe districts controlled by the out-party can be swapped with citizens in marginal districts, who are more likely to vote for the opposition, in hopes of flipping the marginal district.” In effect, in the redistricting game incarcerated persons act like jokers in a deck of cards. Nominal constituents who can be shuffled around between districts for strategic advantage, but whose lack of voting rights denies them any opportunity to influence subsequent election outcomes. This allows them to be deployed in a way that strengthens the hands of the majority party’s candidates while weakening the cards held by their opponents.
“To the extent that such gerrymandering does occur,” Kelly continues, “we should expect to see a movement of prison populations from marginal districts and safe seats held by the out-party to safe districts controlled by the majority party, particularly after a switch in partisan control.” And that’s precisely what the study’s results showed, although there were some inconsistencies in the extent to which line drawers in different states had been prepared, or able, to utilize this tool. The two most aggressive prison gerrymanders were created by the Republicans in Texas and Florida. There, tens of thousands of incarcerated persons were shifted from competitive Democratic districts into safe Republican ones. The net effect of these changes was sufficiently large to flip control of several state senate seats.
In contrast, Democrats in California and Republicans in Michigan, though presented with opportunities similar to those of their southern counterparts, declined to push the envelope. Kelly credits this to a desire among California Democrats to implement a plan that protected existing incumbents, in the hopes of forestalling an effort to create an independent redistricting commission (they were unsuccessful), and stricter legal constraints on the redistricting process in Michigan. Across the nation, in situations where a party took control of the redistricting process when they had not held it during the prior decade, an average of five thousand prisoners per state were shifted into seats that party held. Prison gerrymandering had now become more than a mere historical oddity. It was a new weapon in the gerrymandering arsenal.
—
Why has there not been a concerted legal effort to end prison gerrymandering? After all, the same “one person, one vote” arguments that proved so successful at the Supreme Court in the 1960s are equally applicable to the malapportionment of districts created by the misallocation of prisoners. And the disproportionate burden that prison gerrymandering places on racial minority groups might also lend itself to a challenge under the Voting Rights Act.
Part of the explanation must surely lie in the fact that for most of U.S. history—until the War on Drugs and subsequent profusion of mandatory minimum sentences created the mass incarceration crisis of today—prison gerrymandering wasn’t really that big of a problem. And when it did become one, things happened slowly, like the proverbial frog crouching unsuspectingly in a pot of water on the stove, gradually being brought to a boil. But the truth is that prison gerrymandering also represents the largely ignored and forgotten stepchild of the gerrymander family. It lacks the wildly contorted and misshapen districts that can be easily satirized in the pages of the popular press, and the backing of moneyed interests and aggrieved politically connected plaintiffs that might provide the impetus to spur on litigation. And on the political side, both mass incarceration and prisoner disenfranchisement have to date been largely bipartisan problems, lending neither side much incentive to rock the boat by pushing for reform. But creepingly, agonizingly slowly, things have finally begun to change.
The first big push to end prison gerrymandering came in New York. The charge was led by the Prison Policy Initiative, whose very first published report in their Prison Gerrymandering Project series, released in 2002, had focused on the state. They partnered with the National Voting Rights Institute and Demos, the New York City–based progressive think tank, to file an amicus brief in a 2005 case that was pending before the U.S. Court of Appeals for the Second Circuit.
In that litigation, Jalil Abdul Muntaqim, who had been convicted of the 1971 murder of two New York City police officers and was serving a life sentence, attempted to mount a vote dilution challenge under the VRA to the state’s felon disenfranchisement laws. The lawsuit was eventually dismissed for lack of standing. Muntaqim was unable to convince the court that the proximate cause of his inability to register to vote was actually the state’s policy toward convicted felons, rather than his lack of residency there (he had previously resided in California and expressed an intent to live with family in Georgia if paroled). But this was still the very first case where arguments against prison gerrymandering had been raised before a federal appellate court.
The lawsuit caught the interest of the New York state senator Eric Schneiderman, who that same year introduced a bill proposing to count incarcerated persons at their home addresses for the purposes of redistricting after the 2010 census. Though his bill stalled in committee, momentum was now gradually beginning to build behind the cause. Similar proposals were introduced in Illinois and Texas, again without success. But as the census grew closer, and the sense of urgency began to increase, change belatedly started to happen.
Maryland was the first state to break through the logjam. In April 2010, with the census in full swing, the general assembly passed, and Governor Martin O’Malley signed into law, the No Representation Without Population Act. It required the state to count prisoners as residents of their home districts for the 2011 redistricting. With the stroke of a pen, the city of Baltimore regained almost six thousand inhabitants that the census had counted as residents of correctional facilities elsewhere in the state, a 1 percent population increase over the numbers that the federal government had supplied. Soon thereafter, the residents of Somerset County elected the first African American city commissioner in their history.
Later that same year, New York finally followed suit. In August, the legislature passed the latest version of Schneiderman’s bill, and Governor David Paterson signed it into law. The next month, Delaware enacted a similar law, followed by California in October 2011. Their legislation, however, came too late for the 2011 redistricting cycle and would instead go into effect following the 2020 census.
But almost immediately, the celebratory champagne had to be put back on ice. A group of African American Maryland voters—represented, the Prison Gerrymandering Project alleged, by “Republican Party attorneys” engaged in “a partisan power grab under the guise of an African-American voting rights lawsuit”—filed suit in federal court claiming that the No Representation Without Population Act violated the Constitution. The three-judge federal panel was having none of it. In their unanimous ruling in Fletcher v. Lamone, they declared, “We find no support in the record for this contention.” “Because some correction is better than no correction,” the judges reasoned, “the State’s adjusted data will likewise be more accurate than the information contained in the initial census reports, which does not take prisoners’ community ties into account at all.”
Around the same time, a state court in New York also upheld their law against a lawsuit claiming that it violated the state constitution. And in 2012, without argument or noted dissent, the U.S. Supreme Court summarily affirmed the lower court decision in the Maryland case. With that decision states were finally given free rein, should they choose to do so, to end prison gerrymandering within their own borders.
The last few years have brought even further progress. In 2019, Washington and Nevada passed legislation pledging to count prisoners at their home addresses for the purposes of redistricting after the 2020 census. In early 2020, New Jersey, Colorado, and Virginia did the same, followed by Connecticut and Illinois in 2021, bringing the total number of states that have adopted such laws to eleven, although Illinois’s will not take effect until after the 2030 census. In addition, according to data collected by the Prison Policy Initiative, literally hundreds of individual counties, cities, and towns in almost every state in the nation have taken similar steps to end prison gerrymandering in their local elections. The strength of the momentum behind reforming the counting of incarcerated persons for redistricting has a clear relationship, in my opinion, to the fact that there really aren’t any particularly compelling arguments against it.
Opponents will contend that the last known addresses of some prisoners might not be 100 percent accurate, or that not every individual who is incarcerated will return to their home community after their release. And while both of these things are undoubtedly true, they ignore the fact that this adjusted method of counting prisoners is, at the very least, no less inaccurate than what the Census Bureau already does. And as the court concluded in Fletcher, it is in all likelihood substantially more accurate.
Using the federal census numbers for redrawing districts will always produce such effects. As people are born, die, and move from one location to another, actual district populations will inevitably deviate from the mathematical ideal of population equality on which they were drawn. This is particularly true later in the decade, by which point the census numbers are almost ten years out of date. This is not an indictment of the movement to end prison gerrymandering. It’s an indictment of our entire district-based electoral system.
One of the local jurisdictions that decided to implement reform was Anamosa, Iowa. The election of Danny Young in 2005 had revealed the staggering inequalities in voting-eligible population between the various wards that were used to elect the city council. These revelations prompted Bertha Finn—“a retired journalist and county clerk as well as an accomplished amateur historian,” according to a 2015 obituary—to organize a petition drive to place a referendum on the 2007 ballot to replace the existing ward system with an at-large selection method. Sixty-four percent of Anamosa voters supported the change.
And as for Danny Young, the accidental politician whose one-vote victory had made him the poster child of the prisoner’s dilemma? He finished tenth in the 2009 city council election, held under the at-large system approved by the voters in the earlier referendum. And while, with only six seats up for grabs, this meant that he would lose his incumbent position on the city council, he did at least have the satisfaction of improving significantly on his 2005 showing. This time he received twenty-two votes.