Redistricting Resources for Legislators and Legislative Staff. Julia Jackson. University of Colorado Denver

Transcription

1 Redistricting Resources for Legislators and Legislative Staff 1 Redistricting Resources for Legislators and Legislative Staff Julia Jackson University of Colorado Denver

2 Redistricting Resources for Legislators and Legislative Staff 2 Contents Figures... 3 Tables... 3 Executive Summary... 4 Introduction... 6 Literature Review... 7 Major Legal Requirements... 9 Redistricting Mechanisms...14 Redistricting Staff...19 Redistricting Technology and Data...21 Research Questions...22 Methodology...23 Results...24 Process...24 Staff...26 Technology...31 Resources...34 Discussion and Conclusion...37 References...40 Appendices...43

3 Redistricting Resources for Legislators and Legislative Staff 3 Figures Figure 1: Map of Congressional Redistricting Institutions...16 Figure 2: Map of State Legislative Redistricting Institutions...17 Figure 3: State Deadlines for Completing U.S. Congressional Redistricting...26 Figure 4: State Deadlines for Completing State Legislative Redistricting...26 Figure 5: Redistricting Staff Relative to Number of U.S. Congressional Districts...27 Figure 6: Average Time Staff Spent Working on Redistricting...30 Figure 7: When Preparations for the 2020 Redistricting Cycle Begin...31 Figure 8: Usefulness of and Familiarity with NCSL Resources...35 Tables Table 1: Active Redistricting Litigation in Table 2: Mechanisms for State Legislative Redistricting...15 Table 3: Official Formats of State Legislative Redistricting Plans...25 Table 4: Change in Redistricting Staff Levels Table 5: Redistricting Expertise For Which States Hired Additional Staff...30 Table 6: Redistricting/GIS Software Used in 1990, 2000, and Table 7: Public Input to the Redistricting Process...34 Table 8: Redistricting Resources...36

4 Redistricting Resources for Legislators and Legislative Staff 4 Executive Summary This project is designed to help the National Conference of State Legislators (NCSL) prepare legislators and legislative staff for the 2020 redistricting cycle. We conducted a survey of state redistricting staff, contacting all 50 states and receiving at least partial responses from 47 states (94 percent). The survey data answer critical questions about how the redistricting process is staffed and managed in each state. They also assess the quality of NCSL s existing redistricting materials and help NCSL determine what other materials might be valuable to its member states. Using prior surveys, we were also able to compare staff levels and technology choices with the 1990 and 2000 redistricting cycles. The survey covered the following areas: Overview questions, providing information about the person completing the survey; Redistricting process questions, covering legal requirements, records retention, and pending legislation and initiatives; Redistricting staff questions, covering the quantity and type of staff assistance used in the 2010 redistricting cycle; Redistricting technology questions, covering geographic information system (GIS) vendors and public map submission; and Redistricting resource questions, reviewing NCSL s available resources and offering the opportunity to recommend others. Survey respondents were primarily nonpartisan legislative staff, the vast majority of whom do not work exclusively on redistricting issues. Most respondents worked on the 2010 redistricting cycle. We found that most states either have redistricting commissions or have considered them, that most states spend 6 months to 2 years working on redistricting, and that about half of the states hired additional staff specifically for redistricting. Staffing levels correspond loosely to the number of congressional districts a state has, but there is a good deal of variation. And even though the redistricting technology field has largely consolidated to two vendors, at least

5 Redistricting Resources for Legislators and Legislative Staff 5 half the states will consider switching to a different vendor for the 2020 cycle. Finally, respondents were overall familiar with NCSL s existing redistricting resources and find them useful. Survey findings provide justification for NCSL support in the redistricting process, even though most respondents were confident in their states ability to manage the redistricting process. They suggest a continued interest in the creation of redistricting commissions and an increased interest in online mapping options, both for staff and the public. Respondents expressed a desire for training sessions focused specifically on redistricting staff, with a particular interest in GIS training. They also asked that many of NCSL s existing resources be updated for 2020.

6 Redistricting Resources for Legislators and Legislative Staff 6 Introduction The National Conference of State Legislatures (NCSL) "provides research, technical assistance and opportunities for policymakers to exchange ideas on the most pressing state issues and is an effective and respected advocate for the interests of the states in the American federal system" (National Conference of State Legislatures, 2009). NCSL serves as a resource for legislators and legislative staff on a variety of issues, including elections and redistricting. NCSL s redistricting and elections program has a staff of three, limiting its ability to research any one topic in depth, and in between redistricting cycles, their focus is generally on other issues. The program manager, Wendy Underhill, welcomed the opportunity to work with a student specifically on redistricting. Redistricting comes after reapportionment. The basic law of reapportionment is found in Article I, Section 2 of the U.S. Constitution, as modified by the 14 th Amendment, and it is short: Representatives [ ] shall be apportioned among the several states [ ] according to their respective numbers[.] The actual Enumeration shall be made [ ] within every subsequent term of ten Years, in such manner as they shall by law direct (U.S. Const. art. I, 2). Census Day is April 1, After the Census has completed its survey, each state is granted Representatives based on its share of the population. This is the reapportionment, and the Census Bureau must provide reapportionment data by December 31, It is then up to the states to draw districts for those Representatives. State legislative maps and local jurisdictional maps are also usually drawn at this time. Additional redistricting data must be released by April 1, 2021 (McCully, 2014), after which the states begin redistricting in earnest. This project focuses on redistricting as it relates to the U.S. Congress and state legislatures. Both functions are usually conducted by state legislatures, though some states use a commission process for one or both (National Conference of State Legislatures, 2009). While much has been written about redistricting, there has been very little focus on its administration. As the professional organization for state legislators and legislative staff, NCSL

7 Redistricting Resources for Legislators and Legislative Staff 7 is the ideal organization to report on this. It is also relevant to the field of public administration. Since redistricting generally only occurs every ten years, the legislators and staff involved are often starting from scratch. Term limits and normal staff turnover mean that many people who drew redistricting maps in the 2010 cycle will not be doing so in For example, Legislative Council Staff in Colorado has found that since 1990 when term limits were implemented in the state, state representatives serve, on average, 4.8 years, and state senators serve, on average, 5.4 years (Jackson, 2015). NCSL expects other aspects of redistricting administration to change as well. For one, the available technology is constantly evolving. For another, with the U.S. Supreme Court reaffirming the constitutionality of redistricting commissions ( Arizona State Legislature v. Arizona Independent Redistricting Commission, 2015), additional states will likely consider their adoption. Since 2015 is the midpoint between redistricting rounds, it is an appropriate time to evaluate the 2010 redistricting cycle and the resources that can be prepared for the next. This project seeks to help NCSL prepare legislators and legislative staff for the 2020 redistricting cycle. It uses survey data to answer critical questions about how the redistricting process is staffed and managed in each state. It also assesses the quality of NCSL s existing redistricting materials and makes recommendations for materials to prepare. I begin this report by reviewing the redistricting literature, with a particular focus on three areas: major case law, commissions, and technology and data. I then detail my research questions and methodology and present the results of my research. Finally, I discuss the results and offer a conclusion and recommendations for NCSL. Literature Review There are two common themes in the redistricting literature that are by necessity beyond the scope of this review: the justiciability of political questions and the role of race. I make this distinction because NCSL, along with much of the legislative staff it represents, is a nonpartisan

8 Redistricting Resources for Legislators and Legislative Staff 8 organization. Making a political judgement about redistricting is not a goal of this research. However, a basic understanding of these two themes is necessary to understand the scope of redistricting literature, so I begin by addressing them briefly before moving into my areas of focus. Gerrymandering is a term that comes up frequently when discussing redistricting. The term was first coined in 1812, and it can be defined as the process of redrawing district lines to increase unduly a group s political power (Levitt, 2010, p6). Researchers have identified a number of problems with gerrymandering ( An interstate process perspective on political gerrymandering, 2006). These problems stem from the idea that voters are not being properly represented by their elected representatives, which can occur when a party obtains more seats in a legislative body than its share of the overall vote would otherwise provide (Polsby & Popper, 1991). It is often the result of agency, or entrenchment (Klarman, 1997) problems, where legislators draw lines to protect their own chances of election (McConnell, 2000). This further leads to noncompetitive districts, which can result in a Congress beset with polarization and negative policy outcomes (Cox, 2004). Justiciability. Once the harms of gerrymandering are identified, many scholars turn their attention to the role of the courts in the redistricting process; whether certain questions are justiciable. The articles cited in the above paragraph advocate a robust role of the courts in policing against gerrymandering. On the other side are those who believe that the redistricting process is inherently political. Fuentes-Rohwer (2003) argues that constituents are actually represented well in the American system of two parties and single-member districts. Others conclude, based on statistical analysis of state legislatures before and after redistricting, that redistricting increases a legislature s responsiveness and reduces partisan bias (Gelman & King, 1994). Fuentes-Rohwer notes that when viewed from the district level, rather than the state or national level, a majority of any district will always elect the candidate of their choice. More importantly, when courts intervene, they simply take sides in highly politicized debates

9 Redistricting Resources for Legislators and Legislative Staff 9 (Fuentes-Rohwer, 2003, p538). Rather than policing political problems, deciding redistricting cases makes the courts more political themselves. If the courts are not the appropriate outlet for reducing the harms of redistricting, what is? Scholars frequently suggest changing who does the redistricting. For example, a strategy of reinforcing political competition by taking the process of redistricting out of the hands of partisan officials offers the prospect of realizing our constitutional values (Issacharoff, 2002, p647). I review this possibility in more detail below in my discussion of redistricting commissions. Race and redistricting. One of the primary ways to challenge a redistricting plan in court is over whether it meets the standards of the Voting Rights Act of 1965, discussed in more detail below. A major debate in the literature is over the effects of racial redistricting on minority representation. For example, Cameron, Epstein, & O Halloran (1996) argue that increasing the number of majority-minority districts does not actually improve minority representation in the legislative process. They analyze the relationship between black voting age population in a district and legislators support for minority issues and then simulate different districts and conduct the same analysis. Lublin (1999) rebuts these findings with a point-by-point critique, concluding that racial redistricting remains vital to the election of African Americans (p186). Major Legal Requirements As it stands, both state and federal courts are heavily involved in the redistricting process. Professor Levitt at the Loyola Law School Los Angeles until very recently maintained a website (Levitt, 2015) listing all the congressional and state legislative redistricting cases for the 2010, 2000, 1990, and 1980, which gives an idea of the volume of court involvement. For the 2010 cycle, Levitt identifies 224 total cases filed, 32 of which are still active as of September In only 9 states were there no cases filed, and 4 of those are states with a single congressional representative, meaning no congressional redistricting occurred. Courts actually

10 Redistricting Resources for Legislators and Legislative Staff 10 drew the congressional maps in 9 states and the legislative maps in 6 states. The states still involved in active litigation over redistricting are detailed in Table 1. Table 1: Active Redistricting Litigation in 2015 State U.S. Congressional Map in Litigation State Legislative Map in Litigation AL AZ FL MD NC T VA WY Source: (Levitt, 2015) A surprising number of redistricting cases are heard at the U.S. Supreme Court. One reason for this is the peculiar way in which constitutional challenges to redistricting plans are handled. Federal law requires redistricting cases to be heard by a three-judge district court panel when they involve the constitutionality of apportionment (Manheim, 2013). These cases can then be appealed directly to the Supreme Court, ensuring the high court will take on a number of redistricting cases. follows: Manheim (2013) describes the reasons courts would review a redistricting case as The Fourteenth Amendment requires that each plan (1) comply with the equal representation principle; (2) not purposefully discriminate against racial minorities; (3) not subordinate what the Supreme Court has called traditional race-neutral districting principles to racial considerations unless that subordination can survive strict scrutiny; and (4) in theory, at least, avoid excessive political gerrymandering. Federal statutory law in turn requires that a plan (5) not result in a dilution of minority voting strength (a section 2 claim); (6) in certain jurisdictions, not reduce minority voting strength as compared to prior levels (a section 5 claim); and (7) in congressional races, not use multi-member districts. Depending on the jurisdiction, there also may be restrictions set forth in state law (p ). The major Supreme Court cases on each of these issues have become essential to the redistricting process.

11 Redistricting Resources for Legislators and Legislative Staff 11 Equal population cases. The Supreme Court for decades refused to intervene in redistricting, calling it a political thicket (National Conference of State Legislatures, 2009). This changed in 1962, when the court held in Baker v. Carr that redistricting plans could be challenged over their constitutionality. This led to a series of decisions in the 1960s known as the redistricting cases, culminating with Reynolds v. Sims (1964). The key finding in this case was that districts must be drawn, so that the vote of any citizen is approximately equal in weight to that of any other citizen in the State. This is the principle now familiar in redistricting law of one person, one vote. This standard has been interpreted strictly in the years to follow, with the court ordering states to draw their congressional districts so that they are as close to equal in population as possible, unless the state can justify variation on other grounds (such as respecting municipal boundaries). For congressional districts, the standard is derived from the Constitution s apportionment clause, but the standard for legislative districts is derived from the equal protection clause (U.S. Const., amend. IV, 1). Consequently, the court has allowed greater deviation among state legislative districts, generally up to a 10 percent difference between the largest and the smallest absent another compelling state interest (National Conference of State Legislatures, 2009). Voting Rights Act cases. The Voting Rights Act of 1965 (VRA) has been very influential in redistricting jurisprudence. The two principal sections of the VRA are Section 2 and Section 5: Section 2 was originally a restatement of the 15th Amendment and applies to all jurisdictions. It prohibits any state or political subdivision from imposing a voting qualification or prerequisite to voting or standard, practice or procedure... in a manner which results in the denial or abridgement of the right to vote on account of race or color. Section 5, on the other hand, applies only to certain jurisdictions covered under the act. A jurisdiction covered under Section 5 is required to preclear any changes in its electoral laws, practices or procedures with either the U.S. Department of Justice or the U.S. District Court for the District of Columbia. The two sections work independently of each other. A change that has been precleared under Section 5 still can be challenged under Section 2 (National Conference of State Legislatures, 2009, p51-52).

12 Redistricting Resources for Legislators and Legislative Staff 12 Thornberg v. Gingles (1986) laid out the most specific standards for assessing violations of Section 2. If a minority group wanted to challenge a redistricting plan as diluting their vote, the court in Gingles required them to prove: 1. The group is large enough and located in a sufficiently compact area to allow them to constitute a majority in a district; 2. The group is cohesive in its politics (racially polarized voting); and 3. The white majority usually votes in such a way as to prevail over the minority group s preferred candidate. The decision also stated that the court must review the totality of the circumstances in deciding whether discrimination occurred. The Supreme Court later limited racial redistricting in Shaw v. Reno (1993). It found that such claims require strict scrutiny, and it identified circumstances under which racial redistricting could be found unconstitutional: When majority-minority districts comply with traditional districting principles, and are drawn to redress racially polarized voting, the Court treats them as constitutionally appropriate because [they are] necessary to secure evenhanded treatment. When raceconscious districting goes further, by abandoning the principles typically used to draw other districts, the Court treats race as having been singled out for exceptionally preferential treatment (Pildes, 1997, p2511). Both Shaw and Gingles address Section 2 claims. Shelby County v. Holder (2013), by contrast, indirectly addresses Section 5 preclearance. What the Supreme Court actually did in Shelby County was to declare Section 4 of the VRA unconstitutional. While Section 5 requires preclearance for certain jurisdictions, Section 4 spells out the formula that identified which states and local governments were subject to Section 5 preclearance. In practice, this means that there are no longer any jurisdictions subject to preclearance, unless Congress enacts a new formula (Howe, 2013) or jurisdictions are bailed in based on discriminatory actions, under Section 3. How this will impact the next round of redistricting remains to be seen. Partisan gerrymandering cases. Federal courts have been reluctant to review claims of partisan gerrymandering, but they have at times waded into this arena. In Davis v. Bandemer (1986), the court found that partisan gerrymandering claims are justiciable, and it suggested that

13 Redistricting Resources for Legislators and Legislative Staff 13 the minority party would need to prove discrimination both in intent and effect. It did not, however, establish an effective way to evaluate such claims. In Vieth v. Jubelirer (2004), a fractured court found that there are no judicially discernible and manageable standards for adjudicating partisan gerrymandering claims. In addition to the plurality opinion, the justices issued a concurring opinion and three separate dissents. This suggests that the Bandemer standard had proved unworkable (National Conference of State Legislatures, 2009, p122), but the fragmented opinions left open the possibility of further attempts to adjudicate partisan gerrymandering. In sum, these Supreme Court rulings require states to draw districts with equal populations that allow minority groups to elect representatives of their choice (within reason). They may also prohibit purely partisan mapmaking, but this is less clear. Further, VRA preclearance will likely not exist in the 2020 redistricting cycle. The court will also hear two redistricting cases in 2016, and I briefly preview these cases below. Harris v. Arizona Independent Redistricting Commission (2016) concerns Arizona s state legislative districts. This case came through the process described above heard by a threejudge panel and appealed directly to the Supreme Court. The court agreed to review whether deviations in district size are justified on the basis of partisan advantage, and also whether the deviations are justified by an attempt to obtain VRA preclearance, especially in light of the Shelby County decision (Denniston, 2015). Denniston speculates that this decision is likely to be narrowly tailored to the facts at hand, rather than creating any broad precedent. Evenwel v. Abbott (2016), meanwhile, has the potential to significantly change the redistricting process. Since setting the equal population standard, the Supreme Court has declined to define population, leaving that up to the states. Most states use total population, namely, the census headcount. In Evenwel, though, the Texas legislative redistricting plan was challenged on the grounds that the total population standard results in districts with an uneven distribution of actual voters. This is because total population necessarily includes people who

14 Redistricting Resources for Legislators and Legislative Staff 14 cannot or do not vote: children, convicted felons, non-citizens, and eligible but unregistered voters. The challengers want Texas redistricters to count only voters. There are examples where alternative population bases have been upheld, but no state has ever been required to use a standard other than total population (National Conference of State Legislatures, 2009, p10). If the Supreme Court sides with the challengers, this case would force a big change in how states conduct their redistricting (Denniston, 2015). Redistricting Mechanisms Though federal law governs much of redistricting, it is up to each state to decide how its district lines are drawn. In the 43 states with more than one U.S. congressional district, there are two separate processes one for congressional districts, and one for state legislative districts. In most states, the legislature is responsible for both. Levitt (2015) classifies redistricting institutions in five ways: 1. the legislature alone draws the lines; 2. the legislature draws the lines with the help of an advisory commission; 3. the legislature draws the lines, but a backup commission is available if the legislature fails to pass a plan; 4. a politician commission draws the lines, meaning elected officials may serve as commission members; or 5. an independent commission that excludes elected officials draws the lines. Commissions vary significantly in their composition from state to state. For example, Levitt classifies Oregon as having a backup commission, but in fact the backup is the Secretary of State alone. Iowa is classified as having an advisory commission, but the staff role in Iowa is unique. The advisory body is actually the nonpartisan Legislative Services Agency, which uses statutory criteria to draft redistricting plans, and which can seek guidance from an independent commission. The Iowa legislature then has three chances to accept or reject proposed plans

15 Redistricting Resources for Legislators and Legislative Staff 15 without modification. After three rejections, the legislature can provide its own plan, but this has not occurred since this system began in 1980 (Levitt, 2015). Table 2 provides more detail about the mechanisms used for state legislative redistricting, where there is greater variation in the processes. Figure 1 and Figure 2 are Levitt s maps, applying the above categories to congressional and state legislative redistricting, respectively. State AL AK AZ AR CA CO CT DE FL GA HI ID IL IN IA KS KY LA ME MD MA MI MN MS MO MT NE NV NH NJ NM NY NC ND OH OK Table 2: Mechanisms for State Legislative Redistricting Legislature with Gubernatorial Approval (Veto Power) Legislature as Sole Authority Legislative Process with Backup Commission Legislative Process with Advisory Commission Partisan or Bipartisan Commission Independent Commission

16 Redistricting Resources for Legislators and Legislative Staff 16 Table 2: Mechanisms for State Legislative Redistricting Legislature with Gubernatorial Approval (Veto Power) Legislature as Sole Authority Legislative Process with Backup Commission Legislative Process with Advisory Commission Partisan or Bipartisan Commission Independent Commission State OR PA RI SC SD TN T UT VT VA WA WV WI WY Total Source: Mooney, 2011 Figure 1: Map of Congressional Redistricting Institutions Source: (Levitt, 2015)

17 Redistricting Resources for Legislators and Legislative Staff 17 Figure 2: Map of State Legislative Redistricting Institutions Source: (Levitt, 2015) The use of commissions is much discussed in redistricting literature, mostly for their potential to reduce partisanship in the process. The commissions in Arizona and California are seen as the most independent, and as such have generated the most interest. [C]ollectively they embody elements of almost every redistricting reform idea ever proposed (Cain, 2012, p1812). Arizona s commission was adopted in 2000 and used for the 2000 and 2010 redistricting cycles (Levitt, 2015). Legislative leaders selected members of the Arizona Independent Redistricting Commission from a pool of 25 individuals chosen by a judicial selection commission. The five final commission members included two from each major political party and one independent. The commission was also given criteria for drawing its maps and was required not to base its plans on prior maps or consider incumbents. California s commission was adopted in 2008 (for state legislative maps) and 2010 (for congressional maps) and used for the first time in the 2010 cycle (Levitt, 2015). The Citizens Redistricting Commission is notable for its unique selection process (Cain, 2012). Prospective

18 Redistricting Resources for Legislators and Legislative Staff 18 commissioners had to meet a number of requirements aimed at keeping politics out of the pool. Applicants (36,000 expressed initial interest in 2010) were screened by the California Bureau of Audits, and those who qualified were invited to complete a supplemental application. This group was then whittled down, complete with strikes from legislative leadership, to 36 each of Democrats, Republicans, and independents. Three Democrats, three Republicans, and two independents were chosen by lottery. The 8 then chose the remaining 6. Cain (2012) reports that the commission was diverse with respect to race, gender, and ethnicity, but not with respect to education and class (p1825). In addition to this novel selection process, California s commission was also given specific criteria to be followed and a specific order in which they were to be applied, was subject to extensive public input processes, and was required to have the support of members from each political party and independents to adopt a plan. Cain offers an early analysis of the California and Arizona commissions, finding that they have not eliminated political controversy and partisan suspicions (Cain, 2012, p1812) and therefore, their plans were just as likely to be challenged in court. He notes that in California, the commission s maps were more compact and more competitive than the plans they replaced, but these improvements were modest, and the plans drew criticism from many directions. Cain also points out that by prescribing the numbers of Democrats, Republicans, and independents, the system inherently skews away from proportional representation, as 44 percent of California voters were registered as Democrats and 31 percent Republicans at the time. An in-depth ProPublica report on the California Citizens Redistricting Commission (Pierce & Larson, 2011) concluded that the national Democratic Party and California s Democratic congressional delegation stealthily and successfully manipulated the process in their favor.

19 Redistricting Resources for Legislators and Legislative Staff 19 Redistricting Staff The two articles about the California commission above also cover the role of commission staff, but in general, this is an area to which the literature pays scant attention. The ProPublica report notes that the commission was overwhelmed by the task at hand (Pierce & Larson, 2011, np), with its expert staff overworked and underpaid. Cain also raises concerns about the process of staffing independent commissions. Because staff require certain expertise that the commissioners lack geographic information systems (GIS), statistics, voting rights law they could, in theory, steer commission decisions in a given direction by skewing the advice and options they give the commissioners (Cain, 2012, p1833). Further, most redistricting consultants have worked for one or the other party (Cain, 2012, p1835), which raises suspicions. Cain notes that Arizona s commission in 2011 was dogged by controversy over its staffing decisions. Colorado s state legislative redistricting commission, inaptly named the Colorado Reapportionment Commission, is made up of commissioners appointed by legislative leadership, the Governor, and the chief justice of the Colorado Supreme Court. In reflecting on his commission service, one law professor expresses skepticism of the commission s staff (Nichol, Jr., 2001). He suggests that the commission s staff, currently borrowed from the state s legislative service agencies, should instead be independent from the legislature, writing: Staff members who report the rest of the year to the Speaker of the House or the Majority Leader of the Senate would be idiots to treat all commission members as equals. And our staff members clearly were not idiots. In a process in which commission members are immensely and correctly skeptical of their colleagues' motives, and in which no commission member fully understands the studies, gadgetry, jargon and software that the redistricting process requires, it is a mistake to have a staff beholden to the legislature (p1031). Another professor who served on the Colorado Reapportionment Commission wrote an ebook on his experience, and his comments on staff are limited to two paragraphs, one of which misspells the name of commission s staff director. The staff, he writes, became skilled at getting the right map up on the screen at the right time (Loevy, 2011, p46). In his conclusions,

20 Redistricting Resources for Legislators and Legislative Staff 20 Loevy recommends the commission staff remain under the administrative direction of the Office of Legislative Legal Services (their current configuration), though he suggests highly trained computer experts be hired instead of existing legislative staff (p106). He found that even under the commission structure, the political parties employed computer experts to draw maps that were then submitted by commissioners as their own. It appears that the role of redistricting staff, and the outside perception of that role, is not well understood. Are staff neutral players, mostly responsible for projecting maps on screens? Or do they wield influence on the process as behind-the-scenes experts? The literature does not answer this question. A limited amount of research is available on the role of legislative staff generally, however, and this may or may not be comparable. Recent research on state legislatures has focused on their professionalization (Grossback & Peterson, 2004), and the role of legislative staff is an area of study within this. Some researchers have linked the size and skill level of legislative staff to the effectiveness of legislative leadership (Jewell & Whicker, 1994). Others find that legislators identify staff as a key source of policy information (Gray, 1999; Hammond, 1996). However, DeGregorio (cited in Hammond, 1996, p545) found that U.S. Congressional staff work within the parameters set by committee norms, issue jurisdiction, and leaders expectations and styles. This research suggests that staff role is limited in actually influencing policy outcomes. Hammond (1996) further describes research concluding that state legislative staff influence depends on the technical nature of the issue, how much attention it receives, and whether it is new and unrelated to ongoing legislative concerns (p562). Following this framework, redistricting is certainly technical and new to most of the legislators involved, but it also receives plenty of attention and is an issue of concern to those legislators. Applying these findings to redistricting staff appears to be inconclusive.

21 Redistricting Resources for Legislators and Legislative Staff 21 Redistricting Technology and Data Conventional wisdom about redistricting suggests that improved computing has made gerrymandering easier and more effective. For example, incumbent entrenchment has gotten worse as the computer technology for more exquisite gerrymandering has improved (Issacharoff, 2002, p624), and data collection and computer technology [ ] improvements have enhanced the capacity to gerrymander effectively (Pildes, 1997, p2553). A comprehensive analysis of district compactness, however, found that districts have consistently become more gerrymandered (less compact) over time, and the trend increased significantly beginning in the 1970s (Ansolabehere & Palmer, 2015). This research suggests that the oneperson, one-vote standard and the Voting Rights Act had more to do with reduced compactness than technology. Additionally, an in-depth analysis of the role of technology in the 2000 redistricting cycle concluded that, if redistricting has become more sophisticated in this round over the last, computing technology seems unlikely to have been the primary cause (Altman, MacDonald, & McDonald, 2005, p343). The researchers surveyed states about the computer programs, data, and consultants they used in the 2000 cycle. They note that from 1990 to 2000, computers became cheaper and faster, but the software capabilities did not differ significantly. Like Ansolabehere and Palmer, they found that trends away from compact districts began in the 1960s and 1970s, when computer technology was extremely limited. This particular study is also useful in that its surveys produced results for the 2000 and 1990 cycles that should be comparable to my information about the 2010 cycle. In another analysis, Altman and McDonald note that technological advances have had a notable impact on transparency and public participation in the redistricting process (Altman & McDonald, 2014). This is also an aspect of redistricting technology addressed in my survey. Issacharoff (2002) and Pildes (1997) further assert that improved data has led to more sophisticated gerrymandering. It is also true, though, that census data is limited. The decennial

22 Redistricting Resources for Legislators and Legislative Staff 22 census is based on an actual headcount, but the Census also conducts an ongoing American Community Survey (ACS), which began in 1996 and replaced the long form questionnaire after the 2000 census (US Census Bureau, 2013). The actual headcount data must, by law, be used in reapportionment. In practice, it is also used in redistricting. The ACS measures many more variables than the headcount, but it relies on statistical sampling. Race and ethnicity questions appear on the full census questionnaire, but other questions of interest to redistricters, such as citizenship, are only found on the ACS. This means that the citizenship data is not on par with population data (Persily, 2011, p774); in particular, it cannot accurately estimate citizenship at the census block level, meaning it cannot be relied upon for redistricting. This is something the Supreme Court will have to consider in the Evenwel v. Abbott case. Additionally, Persily points out that, voting rights law in some circuits [ ] demands more information than the census can give (p779). More broadly, there are distinct groups whose allocation into districts raises political questions no matter how they are counted. This includes U.S. citizens residing abroad, college students, military members, and prisoners. Research Questions While the literature paints a good picture of how redistricting is conducted, many questions remain about specific procedures, especially from the perspective of NCSL s members legislators and legislative staff. This project addresses some of these questions. My broad research questions are: How are the congressional and legislative redistricting processes administered in each state? What resources can NCSL provide to help legislators and legislative staff with this process?

23 Redistricting Resources for Legislators and Legislative Staff 23 Methodology To better understand the redistricting process, I surveyed staff members who have worked for their legislatures on redistricting. The survey was sent to NCSL s list of redistricting contacts, drawn from the membership of its Standing Committee on Redistricting and Elections, on September 22, The survey reached all 50 states but in some cases went to elections staff rather than legislative staff. Survey questions were requested by Wendy Underhill at NCSL, and then evaluated to determine whether the desired information could be obtained elsewhere. For example, a series of proposed questions about whether state redistricting plans were challenged in court was left off the survey because Levitt s website provides thorough information about this issue. Many of the survey questions, such as a question about records retention, represent information requested from NCSL by state legislative staff. I prepared the survey and had it reviewed by Professor Ely and Wendy Underhill. Two people with experience staffing redistricting also graciously agreed to review early drafts of the survey. Michelle Davis, an employee of the Maryland Department of Legislative Services and staff co-chair of NCSL s Standing Committee on Redistricting and Elections, and Kate Watkins, an employee of Colorado Legislative Council Staff and staff on the 2011 Colorado Reapportionment Commission, provided valuable feedback. Wendy Underhill did a test run of the survey to make sure the form worked. The survey was prepared in Google Forms, which allows for a variety of question types and free distribution. We also had a Word version available, but only two states used this to actually complete the survey. The areas covered in the survey are: Overview questions, providing information about the person completing the survey; Redistricting process questions, covering legal requirements, records retention, and pending legislation and initiatives; Redistricting staff questions, covering the quantity and type of staff assistance used in the 2010 redistricting cycle;

24 Redistricting Resources for Legislators and Legislative Staff 24 Redistricting technology questions, covering geographic information system (GIS) vendors and public map submission; and Redistricting resource questions, reviewing NCSL s available resources and offering the opportunity to recommend others. A copy of the survey as it appeared to recipients is provided as Appendix A. We received 22 responses within a week of sending the survey, including 12 in the first 24 hours. In the end we collected at least partial responses from 47 states. Results The people responding to the survey were primarily nonpartisan legislative staff (33), which is not surprising since these make up the bulk of standing committee participants. Seven respondents were partisan legislative staff, three work for a state elections office, and four were hired by commissions specifically for redistricting. 37 respondents, or 78.7 percent, staffed the 2010 redistricting cycle. Seven work exclusively on redistricting issues at this time, but only three of those are legislative staff (two nonpartisan, one partisan). Survey results are presented in four categories: process, staffing, technology, and NCSL resources. The survey form did not require responses to any questions, so there are times when some states did not respond to a particular question. In these cases, the non-responses are not tabulated in the results. Process Our process questions included whether states have proposals to create a redistricting commission, requirements for the type of bill adopting a redistricting plan, procedures for correcting technical errors, records retention requirements, and deadlines for completion of the redistricting process. 22 states identified current proposals to adopt some form of redistricting commission. Of the 24 states that did not, 16 already have commissions in place. The remaining eight were Alabama, Delaware, Kansas, Nevada, North Dakota, Rhode Island, Utah, and Wyoming.

25 Redistricting Resources for Legislators and Legislative Staff 25 Delaware, Nevada, and North Dakota noted that they have had proposals in prior years. Therefore, nearly all responding states either have redistricting commissions or have considered them. The official formats of state redistricting plans are summarized in Table 3. The majority of states present their redistricting plans as block assignment files, which are created by GIS software. Metes and bounds bills describe physical features of the local geography, along with direction and distances. Other types identified by respondents tended to use other geographic descriptions all relied on some combination of counties, townships, voting tabulation districts, precincts, tracts, or blocks. Table 3: Official Formats of State Legislative Redistricting Plans Number of Type of Bill States Block Assignment File 25 Metes and Bounds 5 Map 3 Combination of Above Types 3 Other Formats 9 In all but three states (Illinois, Ohio, and Utah), the bill types were the same for legislative and congressional plans. Eleven states said that the bill format is specified in law, and these statutory citations are provided in Appendix B. Thirteen states have an official procedure for correcting technical errors, and statutory citations for these procedures are also provided in Appendix B. A quarter of respondents did not answer the question about their state s redistricting records retention policies, suggesting that they did not know the answer, or said they did not know. Another eight states said they had no specific policy. 15 respondents said that redistricting records fall under their state s general records retention policies, either for state records generally or for legislation specifically. Twelve states identified specific policies for redistricting records, but they generally described these vaguely. No state identified a specific statutory requirement.

26 Redistricting Resources for Legislators and Legislative Staff 26 Figures 3 and 4 graphically represent the deadlines for states to complete their congressional and legislative redistricting, respectively. Many of these deadlines are tied to either legislative session lengths or filing deadlines for 2022 primary elections, but as the figures show, there is a good deal of variation. Almost a year and half separates the first and last congressional deadlines, and the gap is wider for legislative redistricting because the Montana legislature only meets in odd-numbered years, pushing the state s deadline back to AR IL IN NE Figure 3: State Deadlines for Completing U.S. Congressional Redistricting ME MT CA CT ID NM PA HI NJ VA WA MN NC OH UT WY AL KS MD OK TN WI LA MA NY April- May June- July August- September October- November December- January February- March April- May June- July August- September IL IN NE DE ME OR Figure 4: State Deadlines for Completing State Legislative Redistricting AR CA CT ID IA NM OH T CO HI LA NJ ND PA SD VA WA MD MN NC UT WY AL KS OK TN VT WI AK MA NY MT April- May June- July August- September October- November December- January February- March April- May June- July August- September October- November December- January Staff Staffing questions on the survey addressed the number and type of support staff for the redistricting process, the average time they spent on redistricting, and whether and how additional staff were hired. We also asked when staff preparations for the 2020 redistricting

27 Redistricting Resources for Legislators and Legislative Staff 27 cycle will begin, how much of the state s redistricting staff are expected to be new to the process in 2020, and a confidence rating for each state s ability to manage the redistricting process. Staff levels. Survey respondents were asked to estimate the number of full-time equivalents (FTE) working on redistricting in the 2010 cycle and to identify the type of staff they included in their count. Where a range was provided, I used the middle of the range for analysis. Additionally, many staff had other responsibilities, and many respondents noted that their numbers were inexact, so these numbers are not presented as definitive. The mean FTEs across respondents was 9.9, with Pennsylvania, New Jersey, New York, and California reporting over 30. Excluding those four states, Figure 5 compares staffing levels with each state s U.S. congressional districts. The correlation coefficient between these two variables (as calculated in Excel) is 0.63, indicating a slight positive correlation. Figure 5: Redistricting Staff Relative to Number of U.S. Congressional Districts Additional source: Burnett, 2011

28 Redistricting Resources for Legislators and Legislative Staff 28 NCSL also collected staffing data for the 1990 redistricting cycle, and Table 4 compares our 2010 data to the 1990 data where available. Table 4: Change in Redistricting Staff Levels State Percent change Alabama 5 5 0% Alaska Arizona % Arkansas California % Colorado % Connecticut Delaware % Florida Georgia 8 8 0% Hawaii % Idaho % Illinois Indiana % Iowa % Kansas % Kentucky % Louisiana % Maine Maryland % Massachusetts % Michigan Minnesota % Mississippi % Missouri % Montana % Nebraska % Nevada % New Hampshire % New Jersey % New Mexico % New York % North Carolina % North Dakota 2 2 0% Ohio Oklahoma % Oregon % Pennsylvania %

29 Redistricting Resources for Legislators and Legislative Staff 29 Table 4: Change in Redistricting Staff Levels State Percent change Rhode Island South Carolina South Dakota % Tennessee Texas % Utah % Vermont % Virginia % Washington % West Virginia % Wisconsin % Wyoming % Mean % Additional source: Prior NCSL research (found in NCSL files; additional information unavailable) Consistent with what we know about our survey respondents, 35 states said their staffing totals included nonpartisan legislative staff. 17 states included partisan legislative staff, 10 used contract consultants, and 8 included elections or other executive branch staff. Alaska, Arizona, California, Hawaii, Pennsylvania, and Washington mentioned staff hired specifically for a redistricting commission. By contrast, Colorado s commission is staffed exclusively by existing nonpartisan legislative staff. Most states reported spending between six months and two years on redistricting. Only New York, which maintains a permanent redistricting office, and Texas, where litigation is ongoing, reported more than four years. This data is presented graphically in Figure 6.

30 Redistricting Resources for Legislators and Legislative Staff 30 Figure 6: Average Time Staff Spent Working on Redistricting Additional hiring. 28 states said they hired additional staff specifically for redistricting. In 11 of those states, hiring decisions were made by a legislative staff agency. Political parties, leadership, or caucuses made the hiring decisions in nine states. In six states, commissioners or their staff made hiring decisions, and in Wyoming, additional hiring was done by the executive branch. The expertise sought when hiring additional redistricting staff is presented in Table 5. Table 5: Redistricting Expertise For Which States Hired Additional Staff Type States Seeking Number of States GIS/Mapping AK, AZ CA, HI, ID, IL, IA, KS, MS, NE, NV, NM, OH, OK, OR, 17 WY, WA Legal LA, MS, MO, NH, NM, PA 6 Political/Legislative AK, ME, NJ, OK, T, WA 6 Administrative/Organizational/ Clerical/Communications HI, IL, KS, PA 4 Data Analysis CA, MA 2 When asked to speculate as to how many of the people staffing redistricting in 2020 would be doing so for the first time, states were fairly evenly split. Eight thought less than a quarter of their staff would be new, eleven predicted 26 to 50 percent, thirteen predicted 51 to 75 percent, and twelve thought 76 to 100 percent would be new. Despite this expected turnover, states were overall confident that they have the resources and expertise necessary to

31 Redistricting Resources for Legislators and Legislative Staff 31 manage the 2020 redistricting cycle. On a scale of 1 to 5, with 5 being very confident, the mean response was states picked 4 or 5, and only Tennessee selected a score below the midpoint. Preparations. Figure 7 shows when states began or plan to begin preparing for the 2020 redistricting cycle. Where a range was given, this chart shows the earliest point in the range. Again, New York has a permanent redistricting office, and so it is excluded from this chart. Though more than half the states do not plan to begin preparing until 2018 or later, 13 states have either already begun to prepare or will do so in the next year, suggesting that some may be seeking new NCSL resources already. Figure 7: When Preparations for the 2020 Redistricting Cycle Begin OR ID KS KY MS NV VA WA AL MN MT PA T IN TN CA GA IL IA LA MD NE NC UT WY CO DE MA ND OH WI AK HI ME MO NH NJ NM SD AZ VT Technology States were surveyed on the mapping technology they used for redistricting, whether they were considering a new vendor for 2020, and if so, what improvements they would be looking for. We also asked about how states gather public input during the redistricting process. Redistricting software. We identified three primary software programs used to draw redistricting maps in 2010 AutoBound (Citygate), Districting for ArcGIS (Esri), and Maptitude for Redistricting (Caliper) and asked states which vendor they used. 24 states (more than half of respondents) used Maptitude, 15 used AutoBound, and 3 used Districting for ArcGIS. Louisiana reported using a customized ArcGIS application as well as Maptitude, New

32 Redistricting Resources for Legislators and Legislative Staff 32 Hampshire and Texas used in-house programs, and three states were unable to identify the mapping vendor they used. (These three are not included in the total when calculating percentages.) Table 6 compares our data with surveys conducted in 1990 and 2000 (Altman et al., 2005). Note that Altman s surveys only covered congressional redistricting, so states with a single congressional district at the time were excluded from his reports. Table 6: Redistricting/GIS Software Used in 1990, 2000, and Software/Vendor Number of States Percentage of Survey Respondents Number of States Percentage of Survey Respondents Number of States Percentage of Survey Respondents AutoBound % % Districting for ArcGIS/other Esri % % Geodist % Mapinfo 1 2.6% Maptitude/Maptitude for Redistricting % % Plan 90/Plan % 2 4.7% REAPS 2 5.3% Custom % % 3 6.7% None 2 5.3% Additional source: Altman et al., 2005, p337 Particularly notable in this comparison is the decline in the number of states using custom applications, which Altman attributes to the decreasing cost of off-the-shelf GIS software. This chart also shows the decline in the number of vendors participating in this market, from five in 1990 to three in 2000 to three (but practically two) in We also asked whether states were considering switching to a different mapping vendor for the 2020 redistricting cycle, and if so, what improvements they were looking for. Several respondents were unsure of their plans at this time, but 19 said they were considering switching, and 23 said they were not. The yes responses represented 9 of 15 AutoBound users and 10 of 24 Maptitude users, a fairly even split. The most requested improvement (five responses) was for better web-based or online mapping options. Four states reported that they would like improved stability, and three states each said they would seek greater ease of use and more customization options. Other improvements states would like were:

33 Redistricting Resources for Legislators and Legislative Staff 33 More options for designing and publishing maps; More reporting options; Lower cost; Public submission capabilities; Open source technology Automation; and Better customer service. Public input. The survey asked whether states accepted public map submissions. This is another question Altman et al (2005) asked regarding the 1990 and 2000 redistricting cycle. 39 respondents (86.7 percent) accepted public submissions in This compares to 81 percent of states in 2000 and 61.5 percent of states in 1990 (Altman et al, 2005, p339). Most of our respondents did not know how many submissions they received from the public. Three states said they received none. 15 states received between 1 and 10 submissions, and 6 states received between 10 and 90. Seven states reported receiving more than 90 public submissions. Four of those California, New York, Pennsylvania, and Texas were among the five most populous states to respond to the survey. The other three are Arizona, Idaho, and Utah. Arizona s new redistricting commission generated a good deal of public attention, which may have led to the high number of submissions, and both Idaho and Utah offered online tools through which members of the public could easily submit maps. Idaho noted that they used Mapitude Online to provide this service, and Utah developed its site with Esri (Hesterman, 2011). Three states (Delaware, Hawaii, and Kansas) said staff reviewed the public maps for technical and legal compliance before distributing them to legislators or commissioners. Most states that received maps from members of the public submitted them directly to legislators or commissioners for their consideration. A few also posted them publicly. Though we did not ask whether the public submissions influenced decisions, as nonpartisan staff are often unable to

34 Redistricting Resources for Legislators and Legislative Staff 34 answer this kind of question, when asked what was done with the publicly generated maps, only Montana mentioned using them to develop commission proposals. Other methods states used to solicit public input on the redistricting process are detailed in Table 7. Most states conduct public hearings, a process familiar to legislative bodies, and 12 states said public hearings were the only method they used to gather public input. Table 7: Public Input to the Redistricting Process Number of Method of Input States Offering This Option Percent of Respondents Public Hearings % Dedicated Box or Phone Line % Comment Form on a Website % Additionally, seven states reported that they provided public work stations with mapping software where members of the public could create redistricting maps. We did not ask about this specifically, so it is possible more states provided this option. Altman et al (2005) reported that 15 states in 1990 and 18 states in 2000 provided public terminals, web sites, or software for public map creation. Resources NCSL was interested in evaluating its existing redistricting resources and seeking suggestions for future projects. Respondents were asked to rate six existing NCSL redistricting resources: 1. Redistricting Law 2010, also known as the Red Book, is NCSL s comprehensive treatise on redistricting law (p vii). It is updated every ten years by members of NCSL s Redistricting and Elections Standing Committee, as well as NCSL and U.S. Census Bureau staff, and it is available online and in hard copy. 2. The 2010 Redistricting Cases web page ( compiled cases from the 2010 redistricting cycle. It was last updated at the beginning of NCSL maintained similar pages for the 2000 and 1990 cycles.

35 Redistricting Resources for Legislators and Legislative Staff NCSL s main redistricting web page ( includes information on several specific topics and is updated as new NCSL research is completed. 4. Redistricting sessions are offered annually at the Legislative Summit, NCSL s major annual membership convention. 5. Redistricting sessions are also offered at the Capitol Forum, NCSL s annual policy and lobbying meeting. 6. Stand-alone redistricting meetings have been offered in the past specifically for legislative staff working on redistricting. The rating options were: very useful, somewhat useful, not very useful, or not useful at all. Respondents could also indicate that they were unfamiliar with the resource. I coded the ratings on a scale of 1-4, with 1 being not useful at all and 4 being very useful, and then found the mean. The results, along with a tally of those unfamiliar with each resource, are summarized in Figure 8. This shows that respondents overall find all the NCSL resources useful, and nearly all of them are familiar with the resources. The only resource with which more than a quarter of respondents were unfamiliar is the Capitol Forum, which is the smaller of NCSL s two annual meetings. Average Rating Figure 8: Usefulness of and Familiarity with NCSL Resources Number Unfamiliar 1.00 Red Book Standalone Meetings Summit Capitol Forum Cases Page Website 0 Average Rating Number Unfamiliar