Redistricting Developments of the Last Decade—and What's on the Table in This One

The Voting Rights Act

Among the significant voting rights developments of the past decade were the U.S. Supreme Court's decision in Georgia v. Ashcroft, and the Voting Rights Act amendment designed to overrule it. In Georgia v. Ashcroft, the Court found that minority influence districts—those where minority voting-age populations were lowered below majority status, but where minority voters may not be able to elect their candidates of choice, but can play a substantial, if not decisive, role in the electoral process—were acceptable substitutes for majority-minority districts under the non-retrogression test applicable to Section 5 of the Voting Rights Act (Becker 2007). ² In the 2006 reauthorization of the Voting Rights Act, Congress chided the U.S. Supreme Court, stating that it had “misconstrued Congress' original intent…and narrowed the protections afforded by Section 5.” ³ Congress amended Section 5 by adding language specifically forbidding preclearance of any electoral change that has the “purpose of or will have the effect of diminishing of ability” of minorities “to elect their preferred candidates of choice.” ⁴

There are outstanding constitutional questions regarding the implementation of the new Voting Rights Act (Persily 2007). Some clarity was provided in a subsequent case, Bartlett v. Strickland, where the U.S. Supreme Court confirmed that, for a racial minority group to make a successful claim under Section 2 of the Voting Rights Act, that group must constitute at least a majority (i.e., more than 50%) of the voting-age population in the district they seek to create. ⁵ The language of Section 2 that the Court interpreted in Bartlett is consistent with the new language of Section 5, so it is likely that Section 5 will be generally understood to prohibit retrogression as it applies to majority-minority districts only. ⁶ Further, many have interpreted the Bartlett ruling to require that the population base be the citizen-voting-age population, not the voting-age population, a topic I discuss below. If this interpretation is accepted, the change will weigh most heavily on growing Latino populations, where a greater proportion of the population is non-citizen.

In the 2006 reauthorization of the Voting Rights Act, the scope of “purpose” was expanded by Congress to “include any discriminatory purpose.” ⁷ This definition of “purpose” overrules Reno v. Bossier Parish School Board II, ⁸ which interpreted Section 5's purpose prong as limited to changes that had a retrogressive purpose. Under the 2006 amendment, redistricting plans submitted by a jurisdiction for Section 5 review cannot be precleared if an unconstitutionally discriminatory purpose has occurred, even if there is no retrogression of minority representation. It is unclear how this purposeful discrimination standard will be implemented in light of the other amendments to the Voting Rights Act (Pitts 2010).

Another major legal development in the past decade was the Court's interpretation of Section 2 of the Voting Rights Act in LULAC v. Perry, ⁹ which involved the Texas re-redistricting plan. While many commentators focused on partisan gerrymandering allegations, the U.S. Supreme Court rejected plaintiffs' partisan gerrymandering claim, with no agreement by a majority of justices on the legal standard that should govern those claims. A majority of the Court did, however, find a violation of Section 2. The question before the Court was whether the state could protect a Republican incumbent by reducing the Latino population in a district in which they were poised to elect their candidate of choice, while compensating for this change by creating another effective Latino district elsewhere. The U.S. Supreme Court found that Section 2's protection applied to the particular Latino community that was poised to elect its candidate of choice, not to the state at-large, and thus that the state could not substitute one effective minority district for another.

The last voting rights issue that bears mentioning is the constitutionality of Section 5. The Court considered a constitutional challenge to Section 5 in NAMUDNO v. Holder. ¹⁰ However, the thunderous case ended in a whimper when the Supreme Court entirely sidestepped ruling on the constitutionality of Section 5 and instead allowed a local authority to apply for bailout under Section 4 even though many assumed only its parent jurisdiction would be allowed to do so. The Court's surprising interpretation of the statute suggests that it is reluctant to weigh in on the constitutionality of Section 5. Some commentators have expressed the view that Section 5 is in jeopardy because Congress did not adequately justify the twenty-five year extension coverage or its continuation of a four-decade old coverage formula during the 2006 reauthorization. If the Court remains reluctant to rule on the constitutionality of Section 5, then other challenges to the Section 4 coverage formula—such as LaRoque v. Holder and Shelby County v. Holder ¹¹ —are also likely to fizzle, though we can expect more challenges to Section 5 in the years to come. More benignly, Republican-controlled state governments, distrustful of the Obama Administration, are likely to seek preclearance from the U.S. District Court for the District of Columbia instead of the Department of Justice.

State Reforms

A few states have reformed their redistricting processes over the last decade. As explained below, California enacted the most noteworthy reform with the establishment of a citizen commission. Florida adopted new redistricting criteria. Delaware, Maryland and New York now require that prisoners be reallocated to their places of origin for redistricting purposes. A seemingly minor Virginia timing reform may play a significant role in light of a legislative stalemate. In other states, legislatures may yet decide to modify their criteria or processes for redrawing district lines.

The most significant reform occurred in California, the most populous state. After several failed reform attempts, California voters narrowly approved ballot initiatives Proposition 11 in 2008 and Proposition 20 in 2010. Proposition 11 established the California Citizens Redistricting Commission and gave it responsibility for state legislative redistricting, while Proposition 20 extended the commission's power to congressional redistricting. ¹² The fourteen commissioners are chosen through a multistage selection process. ¹³ Five must be registered Democrats, five registered Republicans, and four must be independents or minor party registrants. They must draw districts following defined criteria. ¹⁴ Commission approval of a redistricting plan requires affirmative votes by at least three Democrats, three Republicans and three independent or minor party members. ¹⁵ If the commission fails to reach consensus, a California Supreme Court appointed special master will be enlisted to draw redistricting plans. Any plan may be overturned by voter referendum. There is reason to be optimistic that the commission can reach broad consensus on this difficult task, as a similarly constituted Arizona citizen commission was able to reach agreement last decade. However, there is also reason to be pessimistic, as the Arizona commission was denied preclearance on its first state legislative plan and had to defend its work in state court from state constitutional challenges. ¹⁶

Reformers in Florida took a different route, changing redistricting criteria but not the redistricting process. In 2010, Florida voters approved ballot measures 5 and 6, which adopted new criteria for state legislative and congressional redistricting respectively. An enforcement mechanism is already in place, at least for state legislative redistricting: the Florida Supreme Court is required to review state legislative redistricting plans for constitutional defects. I suspect that the court will focus primarily upon the criterion to “utilize existing political and geographical boundaries” as this provides a bright line test, and less on an “intent to favor or disfavor a political party or an incumbent” as intent is difficult to identify and courts in other states have been reluctant to enforce similar criteria. ¹⁷

“Prison gerrymandering” refers to the problem of padding the populations of districts within which prisons are located with prisoners who are ineligible to vote. ¹⁸ Delaware, Maryland, and New York passed laws that adjust their local and state legislative census population data to reallocate prisoners to the places from which they originated prior to their arrests. ¹⁹ These states will use unadjusted population data for their congressional redistricting. The new laws are similar in character to those of Hawaii and Kansas, which adjust their state legislative redistricting populations for non-resident military personnel and students. ²⁰ Prisoners will be reallocated to their places of origin using prisoner ingest documents. For those states and localities that cannot perform this labor-intensive reallocation or are late to the game, the Census Bureau has released group quarters population counts, which can be used to remove prisoners from the population counts. ²¹

Plaintiffs in the Texas re-redistricting lawsuit challenged the constitutionality of mid-decade redistricting. ²² However, the U.S. Supreme Court found that whether to redraw districts in mid-decade is not a federal constitutional question (Levitt and McDonald 2007). Some states have state constitutional language allowing only one bite at the redistricting apple, thereby forbidding mid-decade redistricting. ²³ In 2004, Virginia voters overwhelmingly approved a change to their state constitution requiring the state legislature to “reapportion the Commonwealth into electoral districts in accordance with this section in the year 2011 and every ten years thereafter.” ²⁴ This innocuous-sounding change may have far-reaching consequences, as the state government is divided between the political parties for the first time during a redistricting. A breakdown of the political process may mean the courts will impose a remedial plan for at least one election, a plan that could remain in effect for the entire decade.

The American Community Survey

In the past, the Census Bureau made available additional statistical data garnered from the census long form well after redistricting was complete. Among these data were detailed characteristics about individuals and households, including their income, education levels, and citizenship status. The census long form is no more. It has been replaced by the American Community Survey (ACS), a monthly rolling survey of the American population that the Census Bureau releases in 1-year, 3-year and 5-year compilations. ²⁵ These data may be useful to establish communities of interest where this is a state requirement. But more importantly, they may be used to measure the citizen-voting-age population, which could be used to determine if a district has a sufficient minority population to be in compliance with Section 2 of the Voting Rights Act in light of the Bartlett decision discussed above. ²⁶

These citizenship data have never been available during redistricting before, and there are a number of methodological considerations that have never been adjudicated. States and localities will likely struggle with how to effectively use these data during redistricting and may find themselves in litigation defending how they constructed them. ²⁷ It is important, then, to understand five important limitations of these data. First, the citizen-voting-age population data available from American Community Survey is created from an aggregation of five years of monthly ACS surveys. The most recent 5-year ACS survey release covers monthly surveys conducted from 2005 to 2009. The ACS is not the census enumeration conducted on April 1, 2010, as the Census Bureau documentation clearly states, “[t]hese estimates will not match counts from the 2010 Census.” ²⁸ Second, the American Community Survey is, as its name clearly states, a survey. As such, it has the same statistical sampling error inherent in all surveys available from the Census Bureau. The statistical error can threaten the validity of a citizenship measure for smaller population districts, such as local and lower chamber state legislative districts. Third, the ACS is weighted. The available citizenship estimates use ACS data weighted to population estimates based on the 2000 census. ²⁹ In some states, there was significant variance between these population estimates and the 2010 census. ³⁰ Fourth, the ACS data are reported in 2000 geography, and a bridge must be constructed between the 2000 and 2010 geography in order to use these data for redistricting. The error that is incurred by this step is unknown. Fifth, the ACS data are reported at the block group level and must be disaggregated to the census block level to be used for redistricting in most states. This step incurs additional unknown error.

Previously, the citizenship data collected in the census long form were released well after redistricting, were aggregated within the current census geography, were weighted to the current census, and had no measure of uncertainty associated with them. The errors inherent in the ACS will likely raise new legal questions. If the standard for an effective minority district under Section 2 is that the minority community must constitute at least fifty percent of the citizen-voting-age population, which estimate should one use, the best guess point estimate without any consideration for error, or should the error be factored in? If so, how? Should the lower bound or upper bound of the error range be used to determine the required minority citizen voting-age population percentage within a district?

Internet-Based Redistricting

In the past ten years, substantial improvements to the communications infrastructure of the United States have made it possible for widespread dissemination of data and computer software through the internet. Ten years ago, persons interested in drawing their own districts were required to purchase relatively expensive commercial geographic information systems, costing thousands of dollars, which was a significant improvement in accessibility over the multi-million dollar systems of two decades ago (Altman, Mac Donald, and McDonald 2005). This round of redistricting promises widespread public access to the redistricting data and tools delivered essentially free to users through the internet. Commercial vendor ESRI is deploying an internet redistricting service that their clients can make available to the public. ³¹ Their claim that they offer “the first and only service that facilitates the creation and sharing of redistricting plans on a web-based platform” is not entirely correct. The Public Mapping Project, led by Micah Altman and myself, with software development by Azavea, had already successfully launched an open-source, web-based internet redistricting application that supported a Virginia Redistricting Competition, the Virginia Governor's advisory redistricting commission, and is (as of this writing) supporting redistricting competitions in Arizona and Michigan. ³² Florida is developing in-house an internet redistricting application. ³³ And there is Dave's Redistricting App, which also enables do-it-yourself redistricting, albeit with limitations. ³⁴

Opening the process has already affected the redistricting policy discourse. Virginia state legislators submitted winning student plans from the Virginia Redistricting Competition as legislative bills, and elements of a congressional plan were incorporated into a recommendation to the state legislature adopted by Governor McDonnell's Independent Bipartisan Advisory Redistricting Commission. This plan also demonstrated a new way of enhancing minority representation in Virginia, and elements there were incorporated into a congressional proposal by the state legislative Black caucus. ³⁵ Further competitions and open-access are planned in other states. Among the most significant may be Florida, where the legislature may find its proposed plans under legal scrutiny by the Florida Supreme Court if someone in the public draws redistricting plans that do better on the state's newly adopted redistricting criteria than the legislature.

Conclusion

Thomas Hofeller, Redistricting Director for the Republican National Committee, describes redistricting consultants as locusts who come out to feed every ten years. ³⁶ The same might be said of the lawyers who work in this area, as redistricting losers frequently seek court intervention. The breadth of legal changes and other developments makes it very likely that redistricting litigation will continue to proliferate in the next decade, providing plenty of food on the table for lawyers and experts alike. ³⁷