“Apportionment” refers to the allocation of representatives in legislative bodies to particular geographical units, while “districting” refers to the design of the geographically based election districts within those units. There is a clear distinction between these processes when it comes to seats in the U.S. House of Representatives. Seats in the U.S. House are first apportioned to states, according to the relative size of each state’s population, through a formula contained in a federal statute (2 U.S.C.S. §2a), while the districts themselves are then designed by the individual states. The distinction between apportionment and districting at the state and local levels is blurred, however. Prior to the adoption of the “one person, one vote” requirement for representational districts at these levels, which mandates that districts be close to equal in population, counties often received specified numbers of representatives in state legislatures, and sometimes municipalities or other preexisting units were allocated seats in county legislatures. Such allocations now almost always conflict with the population equality requirement, resulting in apportionment and districting at these levels becoming essentially one process, performed by the state or locality. The expressions “reapportionment” and “redistricting,” therefore, are used interchangeably at these levels.
The revision of geographical districts to elect legislators in the United States is required after each decennial census. This enumeration of the country’s population happens every year that ends in zero. The new population counts almost always require the revision of U.S. House districts, state legislative districts, and local legislative districts in order to bring them into compliance with the population equality requirement. The purpose of this process, according to the U.S. Supreme Court, is to provide the people with “fair and effective representation” within these bodies (Reynolds v. Sims 1964).
The one person, one vote rule was adopted in the 1960’s, after the U.S. Supreme Court in Baker v. Carr (1962) held that “malapportioned” election arrangements (those with districts with large differences in the size of their population) could be challenged on constitutional grounds. The election districts employed to elect most legislative bodies varied greatly in population at that time. The Court subsequently held during the decade of the 1960’s that the basic one person, one vote rule applied to each state’s districting arrangement for U.S. House elections, to districts for all chambers of state legislatures, and to districts used to elect local governing bodies such as county boards, municipal councils, and local school boards. The constitutional provision requiring this for the U.S. House is Article 1, Section 2, which requires that U.S. representatives be elected “by the people of the several states,” and that for the state and local bodies is the Equal Protection Clause of the Fourteenth Amendment.
Extensive redistricting now occurs at every level of government across the country following the decennial census. One might think, especially given that this process is now computer assisted, that this is a routine administrative task. People are simply removed from districts that are overpopulated, according to the new population counts, and added to the districts that are underpopulated. While this may appear initially to be a simple, straightforward task, it is far from it. The task becomes complex and conflictive because there are numerous ways in which these adjustments can be made, many of which are perceived to have very different political consequences. Different revisions can have different impacts on the electoral fortunes of incumbent legislators. Some may have their reelection prospects enhanced by removing from their district people who are viewed as not likely to vote for them and/or by adding others viewed as likely to support them. Others may find their reelection prospects diminished by the changes. Some may even be placed in the same district as another incumbent, forcing them to compete with each other for reelection. The demographic characteristics of many districts can also be affected by the revisions, resulting in, for example, African American and Latino candidates, and sometimes Native American candidates, having better or worse chances of being elected. And the number of candidates affiliated with a major political party that is likely to be elected, or reelected, can vary greatly depending on which plan is adopted.
The conflictive nature of this task is exacerbated by the fact that the responsibility for deciding which redistricting arrangement will be adopted is almost always placed in the hands of politicians, generally the legislators directly affected by the revisions. State legislators typically adopt the new districts from which they themselves will run for reelection, as do local legislators. State legislators are also typically responsible for adopting the new U.S. House districts for their state, a process that incumbent members of Congress try to influence, often with great success. The ascent of elected governors and local chief executives is almost always required as well before a plan can be implemented. Even when the responsibility for redistricting is placed outside the legislature, it is almost always allocated to elected officials or their representatives. With politicians in charge of something with such immediate political consequences, no one should be surprised if fair and effective representation for the public is not a high priority in their decisions, as opposed to their rhetoric, especially if one party controls both the legislative and executive branches of the government.
The adoption of new districts, however, is often not the end of the process. Those who feel disadvantaged by the results often take their complaints to state or federal courts, with forum shopping to find favorable judges a common practice. The political features of the plans are sometimes attacked directly, with complaints that they are blatant, or subtle but still outrageous, “gerrymanders” (i.e., plans that systematically advantage a particular party’s or a group’s ability to elect representatives of their choice). They are also attacked indirectly through complaints that the districts do not satisfy the one person, one vote rule, or that they do not conform to districting criteria like contiguity and compactness. Litigation often extends the process for several years, occasionally into the later part of the decade and even into the next.
ONE PERSON, ONE VOTE
The basic constraint on districting arrangements is the one person, one vote rule. Any political considerations driving the design of the districts must be achieved within districts that comply with this standard. The standard itself, set by the U.S. Supreme Court, is different depending on whether the districts are used to elect members of the U.S. House or state and local legislative bodies. Larger differences in population are allowed in state and local districts than in congressional districts.
State and local districts are assessed by calculating how far they deviate from the average, or “ideal,” population for all of the districts used to elect the particular elected body. The size of the deviation from the ideal is divided by the ideal and expressed as a percentage of the ideal. (A district that contains 9,500 people when the ideal is 10,000, for example, has a deviation of −500, which divided by 10,000 is −5 percentage points.) If the largest positive and negative deviations within a set of state legislative or local legislative districts, when combined, are 10 percentage points or more, then if the plan is challenged in court, the state or local government must justify the deviations. These justifications must entail “rational” polices that are free of “arbitrariness and discrimination.” State and local plans, therefore, almost always have combined deviations of less than 10 percentage points, in order to avoid having to provide such justifications for the lines.
Districting arrangements that remain within this 10 percentage point standard are “presumptively constitutional” and are usually found to satisfy the one person, one vote rule for state and local districts, but they are not immune from invalidation. Plaintiffs may still challenge the districts on population equality grounds, but in this situation they have the burden of proving that the deviations are arbitrary or discriminatory. In the post-2000 round of redistricting, new districts for the Georgia House of Representatives and Georgia Senate were found by a federal court to violate the one person, one vote rule, despite the fact that the combined deviations in each totaled 9.98 percentage points. The court held that the deviations were not justified because they were the result of systematically underpopulating districts in Democratic areas in Atlanta and the rural parts of the state while overpopulating those in the growing Republican suburbs, and also of protecting Democratic incumbents but not Republican incumbents (Larios v. Cox 2004).
The 10 percent threshold does not apply to congressional districts. No population deviations in congressional districting plans are considered “presumptively constitutional.” If a state’s districts for the U.S. House are challenged in court on one person, one vote grounds, all deviations, no matter how small, have to be justified by the state. The districts in the congressional plans adopted by states therefore are drawn to be extremely close to equipopulous, and deviations in these plans are often expressed in terms of the actual number of people in a district that is above or below the ideal population rather than in percentage terms. Congressional districts in a number of states are either equal in population or as close to equal as possible, given the state population and the number of districts it must be divided among. When Pennsylvania revised its congressional districts following the 2000 census, it adopted a plan in which the largest and smallest districts differed by only nineteen people (from 646,361 to 646,380). This plan was found to violate the one person, one vote standard because those designing the districts decided not to try to lower the deviations after they reached the nineteen-person figure, rather than “zero out” the deviations. A federal court therefore invalidated this plan for not being a “good faith effort” to achieve as low a deviation as possible (Vieth v. Pennsylvania 2002).
Numerous districting plans can comply with the one person, one vote rule and yet be unfair to various political groups. Some racial and language minority groups have special protections against discriminatory districts under the federal Voting Rights Act (43 U.S.C.S. §1973). This federal statute protects African Americans, Hispanics, Asian Americans, Native Americans, and Native Alaskans from districting arrangements that discriminate against them. One section of the act, Section 5, which applies only to specified governmental jurisdictions, primarily southern states and localities, prohibits plans that have an unnecessary “retrogressive” impact on these groups, meaning new plans that are worse for these minorities than the plans being replaced. Another section of the act, Section 2, which applies nationwide, prohibits districting arrangements that dilute the voting strength of these groups. This section prohibits dispersing minority voters across districts, and/or packing them into one or a few districts, in a way that negatively affects their ability to elect representatives of their choice, regardless of what the previous plan may have been like. The Massachusetts House of Representative districts in the Boston area adopted during the post-2000 round of redistricting, for example, were invalidated by a federal court because they conflicted with this provision (Black Political Task Force v. Galvin 2004).
These provisions were widely interpreted by those designing and adopting districts following the 1990 census as requiring the creation of majority-minority districts wherever possible, even if the districts had contorted shapes. This interpretation was rejected by the U.S. Supreme Court, however, which found benign “racial gerrymandering” to be unconstitutional. Districts that violated traditional districting criteria such as contiguity, compactness, and following the boundaries of smaller political subdivisions, if the “predominant purpose” for doing so was racial, would be subject to strict scrutiny under the Equal Protection Clause, an elevated standard rarely satisfied by those challenging a district (see Shaw v. Reno 1993 and Miller v. Johnson 1995). This resulted in many of the majority-minority districts adopted that decade being invalidated and replaced by majority-white districts.
When one party is in control of the districting process, it is common for that party to adopt districts that advantage itself. This phenomenon is theoretically unconstitutional (Davis v. Bandemer 1986), but no districting plan has been invalidated for being a partisan gerrymander. Legal standards for detecting this practice are not agreed upon. The Supreme Court revisited this issue in 2004 in the case Vieth v. Pennsylvania. Democrats in Pennsylvania alleged that the congressional districting plan for that state following the 2000 census, adopted by the Republican-controlled legislature and approved by the Republican governor, was a massive gerrymander that would give Republicans victories in thirteen of the state’s nineteen districts (68.4 percent), despite a relatively even partisan balance among the state’s voters.
The Court was seriously split in its response to this claim. Four justices, concluding that there are no standards for assessing gerrymandering claims, voted to overrule the Bandemer decision and let gerrymandering proceed without any judicial oversight of the practice. Four others upheld the Bandemer determination, but did not agree on an appropriate test, offering among them three different tests for detecting gerrymandering that the plaintiffs could attempt to satisfy. The remaining justice held that it was premature to declare that no manageable standards could be found, but then held against the Pennsylvania plaintiffs. The lack of clarity in the Court’s response offers little guidance for those designing or adopting districting arrangements, and is widely interpreted as leaving the door open for continued, even blatant, gerrymandering.
A new partisan issue arose in the post-2000 round of redistricting. Two states, Texas and Colorado, chose to “re-redistrict” congressional lines after elections were held under new lines in 2002. The districts used in 2002 had, in each state, been adopted by courts after the state legislatures could not agree on new districts. The Republicans gained control of the redistricting process in subsequent state elections, and proceeded to adopt a second plan more favorable to their party. While re-redistricting required by the judicial invalidation of plans has been common, re-redistricting not required by a court has not been. The partisan motivations in redrawing the district lines a second time were obvious, stimulating concerns that partisan gerrymandering could become a common practice following state and congressional elections. Colorado’s second revision was invalidated by the state’s supreme court as a violation of the state’s constitution (Salazar v. Davidson 2003). The Texas plan, approved by a federal district court, was vacated by the U.S. Supreme Court in 2004 (Jackson v. Perry) and remanded to the district court.
Reapportionment and redistricting in the United States are supposed to guarantee that the representation of the people will be fair and effective. The process is now widely criticized, however, as one through which elections are frequently rigged. Districts may be close to equal in population, but still far from fair in the eyes of many. Redistricting is now often viewed as an embarrassing dimension of American democracy. Given the inability of the Supreme Court to protect voters against blatant political gerrymanders, it is often proposed that the process be taken out of the hands of politicians and placed with politically independent boundary commissions, as is done in Australia, Canada, and the United Kingdom. Other proposals entail changing the manner in which elections are conducted. These call for replacing the winner-take-all districts so prevalent today, in which the design of districts is so critical to how votes convert into seats for political parties and other groups, with semi-proportional election systems such as limited, cumulative, and preference voting within multi-seat districts, which can greatly reduce the political consequences of where district lines are placed. The debate over how to improve this part of American democracy is no longer limited to arguments about population equality, contiguity, and compactness, but concerns more fundamental restructuring of the redistricting process and election systems themselves.
Robert G. Dixon Jr., Democratic Representation (New York: Oxford University Press, 1968); Richard L. Engstrom, “The Post-2000 Round of Redistricting,” Publius 32 (Fall 2002): 51–70; Bernard Grofman, “Criteria for Districting,” UCLA Law Review 33 (October 1985): 77–184; and Michael P. McDonald, A Comparative Analysis of the 2001 U.S. Redistricting Institutions, special issue of State Politics and Policy Quarterly 4 (Winter 2004).