Institution-Selecting Regulations

Under Article I § 4 of the U.S. Constitution, states are primarily responsible for determining their congressional and state legislative district boundaries, however, superseding federal criteria apply. In this two-layered process, state constitutions and statutes govern who draws district lines, when, and how, but if states do not satisfy federal or state criteria, courts may intervene.

Until recently, political scientists adopted weakly valid measures of gerrymandering, coding partisan or bipartisan gerrymanders based on which party controlled the state legislature and governor's office (e.g., Erikson 1972; Abramowitz 1983; Born 1985; Niemi and Winsky 1992). Since redistricting does not always follow the textbook legislative process, which itself is susceptible to party factionalism, a more valid classification of gerrymandering types examines the motivations behind a map (e.g., Ayres and Whiteman 1984; Basehart and Comer 1991; Cain and Campagna 1987; Campagna and Grofman 1990; Gelman and King 1994b). Understanding how redistricting institutions structure outcomes follows (McDonald 2004).

There are two basic types of redistricting institutions, those that follow the legislative process and those that use a commission at some stage. Twenty states empower commissions as the sole congressional or state legislative redistricting authority, a source of proposals to the legislature, or a backup if the legislative process fails. Circumstances outside the state's normal process might produce a different outcome than otherwise predicted, such as bipartisan maps adopted where one party controlled the process post-2000 in California and Ohio (see McDonald 2004; for post-1980 examples, see Ayres and Whiteman 1984, 310). Some states do not even use the same process for congressional and state legislative redistricting.

A redistricting commission may function differently from a legislature, but for a few exceptions these institutions are designed to concentrate power in party leaders, not to diminish it. There are two general types of redistricting commissions, partisan and bipartisan. On a partisan commission an unequal number of partisan appointees adopt a map by majority rule, leading to a partisan gerrymander. On a bipartisan commission an equal number of partisan appointees adopt a map or select a tiebreaking member by a supermajority rule. Both cases, either the selection of the map itself or the negotiations involved in selecting a tiebreaking member, lead to a bipartisan gerrymander. Commissioners are often either elected officials or handpicked lieutenants. Thus, these commissions tend to produce gerrymanders similar to those produced by the legislative process, if not more effective ones, because potentially meddlesome legislative backbenchers are not involved.

Ironically, good government groups advocate commissions as the best method for redistricting reform. These groups successfully adopted commissions by way of ballot initiatives in Washington in 1983, Idaho in 1994, and Arizona in 2000.¹

The archetype is Arizona's Independent Redistricting Commission (AIRC), which is distinguished from other states' highly political commissions in that it is designed to promote independence of commissioners, adhere to a defined set of criteria, be transparent, and provide mechanisms for public input.² The AIRC blends Cox's three reform types. It is an institutional departure from the legislative process and it operates under process-based criteria, one of which is outcome-oriented: drawing competitive districts. One size does not fit all, as reformers propose similar institutional structures that best fit their state's politics.

Reform groups do not advocate for Iowa's process, which is often lauded for the electoral competition it produces. Iowa's institution is similar to non-partisan election commissions used in other nations (Handley 2005), in that non-partisan legislative staff sends maps drawn under a well-defined set of criteria to the state legislature for approval. But there are reasons to doubt the model's portability to states where legislative staff is politicized. Furthermore, the state's competitive elections are a function of Iowa's overall competitive landscape, and Iowa does not have the complicating factor of drawing minority-majority districts. Finally, an escape clause permits Iowa's legislature to take control, as it did post-1980 (Ayres and Whiteman 1984).

Reformers' goal of commissioner independence is endeavored through qualifications (McDonald 2006, 237–8). The 1960 Alaska constitution innovated this approach by requiring that “none … may be public employees or officials.”³

Hawaii and Missouri commissioners are forbidden from running for office in the districts they draw.⁴ These two approaches were subsequently combined in Idaho and Alaska reforms.⁵ Prospective AIRC members must abide by these combined qualifications and are further vetted by the state's Commission on Appellate Court Appointments.⁶

Spatial modeling predicts pivotal commissioners' preferences play an important role in expected outcomes. Commissions with an equal number of partisan appointees using majority rule may break deadlocks by the appointment of a tie-breaking member, either selected by a commission majority vote or an outside authority. While the former tiebreaker choice may produce a bipartisan gerrymander, the latter may lead elsewhere. In 2000, the New Jersey Supreme Court appointed Princeton political science professor Larry Bartels as the tie-breaking member of the state's legislative commission. Bartels had partisan members bid for his vote by stating he would support the map scoring best on his proposed criteria. In a less direct way, the AIRC seeks a similar outcome by forbidding the tie-breaking member from being registered with a major political party.

Regardless of how districts are drawn, courts may have a final say. Redistricting is often conducted under severe time constraints, as maps must be adopted prior to candidate primary election filing deadlines. A constitutionally defective map may remain in place when there is not enough time to litigate constitutional violations. Court review can be expedited by formally requiring courts to review maps before enactment or by providing state supreme courts with original jurisdiction on redistricting matters.

Process-based Regulations

Process-based regulations seek to remove discretion from a redistricting authority through adherence to criteria. Arizona's commission must comply with a comprehensive criteria list that ironically worked against reformer's goals in that state (McDonald 2006). Where Arizona's constitutional language permits some commission discretion, Ohio's Issue 4, rejected by voters in 2005, would have gone further by requiring a proposed commission to automatically adopt the submitted map that scored best on constitutionally codified criteria. Some have even called for programming a computer to draw districts, though the technology has yet to advance where true multiple-criteria optimization is possible (Altman, Mac Donald, and McDonald 2005). The danger in adopting seemingly neutral criteria is that they may have subtle “second-order” biases (Parker 1990) that intentionally or unintentionally affect a political outcome.

State redistricting criteria are constrained by superseding federal criteria. In the wake of Baker v. Carr, districts must be of relatively equal population. Practically, this means that redistricting must occur following the census at the start of each decade.⁷

The Supreme Court sets different standards for congressional and state legislative districts, generally permitting a 1% population deviation between the largest and smallest congressional district and a 10% deviation among state legislative districts.⁸ Federal law requires single-member congressional districts,⁹ and state law is controlling for state legislative districts. States must respect the Voting Rights Act, which I discuss below as an outcome-based regulation.

Cox (2004, 756) argues, “There is general pessimism about the ability of process-based regulations … to thwart partisan gerrymandering efforts.” The pessimism arises from application of the equal population standard in the 1960s, which was predicted to significantly constrain gerrymandering (e.g., Cox and Katz 2002; Musgrove 1977; White and Thomas 1964). The ubiquitous scholarly redistricting research since indicates this goal was not realized. As Gelman and King (1994b, 553) note, “population equality guarantees almost no form of fairness beyond numerical equality of population.” Perhaps practice makes perfect, as redistricting authorities learned to effectively gerrymander when repeatedly required to draw districts (Niemi and Winsky 1992, 566).

Cain, Mac Donald, and McDonald (2005, 16) herald the current period of redistricting litigation as the “non-federal criteria period.” While some states have long-standing criteria, successful pressure for reform and state adjustments to evolving federal race standards prompted states to adopt more, sometimes complex, criteria. Litigants have thus increasingly found state courts a fruitful venue to challenge a map.

There are a number of criteria that vary among the states. Some states codify equal population standards that are more constraining than the 10% deviation permitted for state legislative districts. Although a recent Supreme Court decision permits mid-decade congressional redistricting under federal law, state law may forbid this activity and state courts have invalidated maps drawn mid-decade (Levitt and McDonald 2007).¹⁰

The relatively non-controversial contiguity criterion is found in state law, though there are examples of districts of questionable contiguity that stretch across water or are connected at a point (Altman 1998). States may require districts to follow existing political boundaries, such as counties, cities, townships, precincts, and wards, which can have odd-shapes or even be non-contiguous. For example, Wisconsin's 61st Assembly District violates contiguity because it contains a non-contiguous ward, but not under state law, which simply defines all wards as contiguous.¹¹

On compactness, academics have muddied the waters, not provided clarity. There are a surprisingly large number of proposed district compactness measures (Niemi et al. 1990), some of which are dependent on a district's orientation (Altman 1998). Iowa codifies a “perimeter to area” compactness standard,¹²

while other states are silent. Without a concrete standard, in the Shaw v. Reno cases the U.S. Supreme Court struck down racially motivated districts on their “bizarre” shape, although shape is not a limiting factor in achieving other political goals (McDonald 1998).¹³ However, state courts have on occasion overturned districts that violated state compactness standards.¹⁴ Conventional wisdom holds that compactness standards favor Republicans (e.g., Cox 2004), though some argue otherwise (Polsby and Popper 1993), and one of the few articles on the topic finds compactness standards reduce minority representation (Barabas and Jerit 2004).

A recent redistricting criteria innovation is respecting “communities of interest.” In response to the Shaw line of court cases that rejected maps based on predominant racial intent, states conceived communities of interest as a euphemism for racial communities, one that would not necessarily trigger court intervention since odd shapes could be justified on a non-racial basis (Cain, Mac Donald, and McDonald 2005). The difficulty with communities of interest is that there is no objective definition. Logically, it must apply to all communities, not just racial communities, opening the nebulous standard as a rationale for achieving any political goal.

The difficulty in defining communities of interest reveals a larger problem with process-based regulation and why there is much pessimism that process-based regulation is in itself insufficient to reform redistricting. In all but the most egregious violations of criteria, courts are deferential to the political process, particularly as it plays out in state legislatures. For example, a Michigan court declared that the redistricting criteria adopted in a previous legislative session were non-binding on the legislature that drew the maps.¹⁵

As a result, criteria often only receive lip service. Further, there is a paucity of research that investigates the political effects of state criteria, so there is little guidance political science can provide to inform the reform debate. A complicating factor is that process-based political effects are heavily contingent on the political geography of a state: geographically-objective criteria may uniquely interact with the distribution of partisans and racial minorities. For this reason, perhaps the most promising avenue for studying effects is drawing maps by simulation under alternative criteria and observing the results.¹⁶

Process-based regulation may be most successful when implemented with institution-selecting regulation, i.e., a redistricting commission that operates under strict criteria. Recent court decisions suggest that state courts will hold commissions to a higher standard than they do legislatures. For example, court cases in Alaska and Idaho required these states' redistricting commissions to address constitutional defects.¹⁷

The challenge, then, is to formulate concrete guidelines that a commission can follow and for which a court can hold a commission accountable.

Outcome-based Regulations

Where process-based regulations seek to limit gerrymandering by removing a redistricting authority's discretion, outcome-based regulations require a political outcome, be it safeguards against dilution of minority votes or overturning of an overtly political map. The federal courts have energetically enforced regulation of racial gerrymandering, but have been less enthusiastic to formulate, much less enforce, other political regulation. As such, legal scholars are pessimistic federal courts will enforce outcome-based regulations other than race (e.g., Issacharoff and Karlan 2004).

The 2006 reauthorization of the Voting Rights Act signals another 25 years of race-based regulation of redistricting, barring a successful challenge to the Act's constitutionality.¹⁸

Here, the courts and Justice Department provide meaningful output-based regulation of redistricting. There is wide consensus in the profession that when racially polarized voting is present, grouping minority communities within or dividing them among districts affects the ability of minorities to elect, in legal parlance, their “candidates of choice” (Brace, Grofman, Handley, and Niemi 1988). The debate is over the effective percentage of minorities to place into districts to maximize their substantive representation (e.g., Cameron, Epstein, and O'Halloran 1996; Lublin 1999) and the policy implications of drawing such districts (Thernstrom 1987; Swain 1995).

The courts are often just another venue for political warfare, even concerning race-related litigation. Since 1970, only 10 of 35 court-approved congressional maps approved prior to the election following the census release did not favor a party or incumbents (McDonald 2006, 230). While political scientists have proposed partisan gerrymandering standards based on minimizing partisan bias, or more generally on maximizing partisan symmetry (Grofman and King 2007), outcome-based regulation of partisan gerrymandering by the U.S. Supreme Court is unlikely to emerge soon. The recent LULAC v. Perry decision reveals that Justice Anthony M. Kennedy, the swing vote on partisan gerrymandering claims, believes partisan gerrymandering is justiciable, but has yet to find a standard to his liking.

Perhaps meaningful intervention can come from state courts, where outcome-based regulation is found in some state laws. Idaho, Iowa, and Washington use similar wording to Hawaii's constitution, which requires that, “No district shall be so drawn as to unduly favor a person or faction.”¹⁹

Arizona, Washington, and Wisconsin (state legislative only), require competitive districts (McDonald 2006).²⁰ Arizona and Iowa operationalize this mandate through process-based regulations forbidding their redistricting authority access to election data and knowledge of where incumbents live.²¹ While enforcement of outcome-based regulation by state courts is possible, ongoing litigation in Arizona over the commission's competition criterion suggests that such enforcement, if indeed warranted, may be elusive.²²