On August 3, 2010, the New York State Senate approved legislation (1) to end the practice known as "prison-based gerrymandering," (2) making it the third state--along with Maryland (3) and Delaware (4)--to eliminate this potential source of liability in the next redistricting cycle. Meanwhile, however, all other states (and many localities) continue to count incarcerated persons at their places of confinement rather than at their home addresses during redistricting, a practice that artificially inflates the population count and the concomitant political representation of the districts where prisons and jails are located.
As several commentators have noted, the result is a distortion of the distribution of political representation in a manner that violates fundamental principles of equality in the democratic process. (5) The communities that suffer most from prison-based gerrymandering are undoubtedly urban communities of color, whose members are disproportionately represented in the incarcerated population. But they are not the only victims of this practice; collectively, all individuals living outside of prison districts suffer a proportionate dilution in voting power. Moreover, because it enhances the political power of districts that house prisons, prison-based gerrymandering has real-world policy consequences by incentivizing opposition to criminal justice reforms that would decrease reliance on mass incarceration. (6)
This Article argues that, in addition to being fundamentally unfair, prison-based gerrymandering also exposes state and local governments to liability in the coming redistricting cycle. Although this Article examines prison-based gerrymandering at the national level, I rely on data from New York to illustrate the democracy-distorting effects of prison-based gerrymandering, and conclude that the New York State Legislature was wise to end prison-based gerrymandering before the current redistricting cycle.
This Article proceeds in five parts. In Part I, I describe the prison-based gerrymandering phenomenon and explain its impact on the political process. The incarcerated population in the United States has exploded in the past half-century, rising to over two million, (7) and standing approximately equal to that of our three smallest states combined. (8) Thus, the location at which incarcerated persons are counted during redistricting, which for much of the nation's history may have been nothing more than a curiosity for demographers, has now become an issue of vital importance for the shape of our democracy.
In Part II, I examine the rationale for counting incarcerated persons at their home addresses rather than at their places of confinement. Most tellingly, nearly all states have constitutional provisions or statutes expressly providing that a person does not gain or lose legal residence by virtue of being incarcerated, which comports with the general legal principle that a person's domicile is the place where a person lives voluntarily and intends to remain. This nearly universal legal rule is consistent with the understanding that a person does not become a constituent of a place simply by being incarcerated there.
In Part III, I address the principal counterargument for maintaining the prison-based gerrymandering system: namely, that individuals should always be counted where they are physically present. While physical presence is undeniably related to the concept of legal residence, this argument ignores the fact that neither state law nor the Census Bureau has ever treated physical presence as determinative of residence; overseas federal employees and military personnel, for instance, are counted in the states that they consider home, even though they are not physically present. And while incarcerated individuals are frequently compared to groups like students or domestic military personnel, who are and should be counted where they are physically present, there are many differences that set incarcerated persons apart.
In Part IV, I turn to legal issues related to the counting of incarcerated populations during the redistricting process. I conclude that states and localities that fail to allocate incarcerated persons to their home addresses during the redistricting process could be vulnerable in at least two ways: (1) for violation of the "one person, one vote" principle of the Equal Protection Clause; and (2) for minority vote dilution under section 2 of the Voting Rights Act of 1965 (VRA). (9) The redistricting process is almost always fraught with litigation, (10) and thus, states and localities would be wise to eliminate a potential source of liability by not engaging in prison-based gerrymandering during the current redistricting cycle.
Finally, in Part V, I address various technical issues connected with ending prison-based gerrymandering, including whether incarcerated persons should simply be excluded from the redistricting population base altogether, rather than reallocated to their home addresses.
In sum, state and local governments that continue to engage in prison-based gerrymandering may be exposed to additional liability during the already litigious redistricting process. Fortunately, states and localities have an opportunity to join Maryland, New York, and Delaware by refusing to engage in prison-based gerrymandering during the current redistricting cycle. Such a move would comport not only with basic legal rules of residence and domicile, but also with broader principles of fairness and equality in the democratic process.
THE PRISON-BASED GERRYMANDERING PHENOMENON
To understand the magnitude of the prison-based gerrymandering problem, it must be placed within the context of incarceration trends over the past several decades. The 2000 Census counted the number of incarcerated persons in the United States at approximately 1.99 million; (11) more recent statistics place the number at about 2.3 million. (12) These numbers represent an explosion from just a few decades ago. For instance, the state prison population, (13) which was approximately 218,000 in 1974, grew to over 1.3 million in 2000. (14) Over the last twenty years, the number of state prisons has grown from fewer than 600 to nearly 1000. (15) Today, the total incarcerated population of the United States is roughly equal to our fourth-largest city (Houston); it is larger than that of fifteen individual states, and larger than the three smallest states combined. (16) If the incarcerated population could form its own state, it would have qualified for five votes in the Electoral College after the 2000 reapportionment. (17)
These trends are mirrored in New York. From 1980 to 2000, the number of prisons in New York more than doubled, from thirty to sixty-five facilities. (18) The rate of imprisonment in New York has grown from 123 out of 100,000 citizens in 1980 to three times that rate in 2000. (19)
As we shall see below, fairness in the democratic process is a casualty of our policies of mass incarceration.
The Census Bureau's Method of Counting Incarcerated Persons
State and local governments generally rely on data compiled by the Census Bureau when drawing election district lines. The Bureau produces several data sets for this purpose, including one data set created pursuant to Public Law 94-171 (the P.L. 94-171 data file). (20) In compiling the P.L. 94-171 data file, the Bureau enumerates most individuals according to the "usual residence rule"; that is, it allocates each person to "the place where a person lives and sleeps most of the time" as of Census Day. (21) The Bureau classifies certain living arrangements as "group quarters" (GQs), such as military barracks, dormitories, and prisons. (22) Individuals living in GQs pose a special challenge for the application of the usual residence rule, because such individuals frequently consider their "homes" to be someplace other than where they usually live and sleep. Nevertheless, the Bureau's P.L. 94-171 data file counts GQ residents where they sleep, which, for incarcerated persons, is at the correctional institutions where they are confined.
The P.L. 94-171 data file does not distinguish between who is a GQ resident and who is not. In other words, states and localities relying on the data file have no way of knowing which individuals are incarcerated and which are not. The result is a distorted view of actual population and demographic trends. As tabulated by the Census, there are more than twenty counties in the United States where more than one-fifth of the population is actually comprised of prisoners. (23)
The current redistricting cycle, however, presents states and localities with a unique opportunity to count incarcerated individuals differently, as the Census Bureau has announced that it will, for the first time, release GQ data in time for state and local governments to utilize that information during the redistricting cycle. (24) As explained in a statement by Census Bureau Director Robert Groves, the Bureau is releasing this data for the express purpose of giving states and localities the ability to reallocate incarcerated populations. (25) Thus, states and localities that rely exclusively on Census data when redistricting will have a new opportunity to correct for the miscount of incarcerated persons.
Distortions Caused By Census Bureau's Rules on Counting Incarcerated Persons
As discussed in greater detail in Subpart IV.B below, the "one person, one vote" principle of the Equal Protection Clause requires that election districts contain roughly the same number of people so that everyone is represented equally in the political process. As long as states and localities continue to rely on the raw P.L. 94-171 data file when drawing election districts, however, the redistricting process will remain distorted by the inflation of population numbers in electoral districts where prisons are located. (26)...