IN THE UNITED STATES DISTRICT COURT FOR THE WESTERN DISTRICT OF WISCONSIN ONE WISCONSIN INSTITUTE, INC., CITIZEN ACTION OF WISCONSIN EDUCATION FUND, INC., RENEE M. GAGNER, ANITA JOHNSON, CODY R. NELSON, JENNIFER S. TASSE, SCOTT T. TRINDL, MICHAEL R. WILDER, JOHNNY M. RANDLE, DAVID WALKER, DAVID APONTE, and CASSANDRA M. SILAS, Plaintiffs, FINDINGS OF FACT & CONCLUSIONS OF LAW v. 15-cv-324-jdp MARK L. THOMSEN, ANN S. JACOBS, BEVERLY R. GILL, JULIE M. GLANCEY, STEVE KING, DON M. MILLS, MICHAEL HAAS, MARK GOTTLIEB, and KRISTINA BOARDMAN, all in their official capacities, Defendants. Mrs. Smith has lived in Milwaukee since 2003.1 She was born at home, in Missouri, in 1916. In her long life she has survived two husbands, and she has left many of the typical traces of her life in public records. But, like many older African Americans born in the South, she does not have a birth certificate or other documents that would definitively prove her date and place of birth. After Wisconsin’s voter ID law took effect, she needed a photo ID to vote. So she entered the ID Petition Process (IDPP) at the Wisconsin Department of Motor Vehicles (DMV) to get a Wisconsin ID. DMV employees were able to find Mrs. Smith’s record in the 1930 census, but despite their sustained efforts, they could not link Mrs. Smith 1 “Mrs. Smith” is not her real name, which I withhold to protect her privacy. The record of her interaction with the DMV is PX421. to a Missouri birth record, so they did not issue her a Wisconsin ID. She is unquestionably a qualified Wisconsin elector, and yet she could not vote in 2016. Because she was born in the South, barely 50 years after slavery, her story is particularly compelling. But it is not unique: Mrs. Smith is one of about 100 qualified electors who tried to but could not obtain a Wisconsin ID for the April 2016 primary. Wisconsin’s voter ID law is part of 2011 Wis. Act 23, enacted the year after Wisconsin Republicans won the governorship and majorities in both houses of the legislature. Act 23 was the first of eight laws enacted over the next four years that transformed Wisconsin’s election system. Plaintiffs in this case challenge the voter ID law, the IDPP, and more than a dozen other provisions in these new laws, none of which make voting easier for anyone. Plaintiffs contend that the new voting requirements and restrictions were driven by partisan objectives rather than by any legitimate concern for election integrity, that these laws unduly burden the right to vote, and that they discriminate against minorities, Democrats, and the young. Plaintiffs contend that the new election laws violate the First, Fourteenth, Fifteenth, and Twenty-Sixth Amendments to the Constitution, and § 2 of the Voting Rights Act. This case was tried to the court in May. Over nine extended days, the court heard the testimony of 45 live witnesses, including six experts, with additional witnesses presented by deposition. The parties submitted lengthy post-trial briefs, and the court heard closing arguments on June 30. The opinion that follows is the court’s verdict. It sets out in detail the facts that the court finds and the legal conclusions that the court draws from those facts. Because of the large number of claims asserted in this case, and the volume of evidence submitted, the opinion is necessarily long, and few readers will endure to the end. But I will 2 try, in a few pages of introduction, to explain succinctly the court’s essential holdings and the reasons for them. I start with a word about my role. It is not the job of a federal judge to decide whether a state’s laws are wise, and I certainly do not have free-floating authority to rewrite Wisconsin’s election laws. My task here is the more limited one of pointing out where Wisconsin’s election laws cross constitutional boundaries. The Constitution leaves important decisions about election administration to the states. But election laws inevitably bear on the fundamental right to vote, so constitutional principles come into play. The standards that I must apply to plaintiffs’ claims require me to examine carefully the purposes behind these laws, and sometimes to draw inferences about the motives of the lawmakers who enacted them. I conclude that some of these laws cannot stand. Wisconsin’s voter ID law has been challenged as unconstitutional before, in both federal and state court. In the federal case, Frank v. Walker, the Seventh Circuit held that Wisconsin’s voter ID law is similar, in all the ways that matter, to Indiana’s voter ID law, which the United States Supreme Court upheld in Crawford v. Marion County Election Board. The important takeaways from Frank and Crawford are: (1) voter ID laws protect the integrity of elections and thereby engender confidence in the electoral process; (2) the vast majority of citizens have qualifying photo IDs, or could get one with reasonable effort; and (3) even if some people would have trouble getting an ID, and even if those people tend to be minorities, voter ID laws are not facially unconstitutional. I am bound to follow Frank and Crawford, so plaintiffs’ effort to get me to toss out the whole voter ID law fails. If it were within my purview, I would reevaluate Frank and Crawford, but not because I would necessarily reach a different conclusion. A well-conceived and carefully implemented 3 voter ID law can protect the integrity of elections without unduly impeding participation in elections. But the rationale of these cases should be reexamined. The evidence in this case casts doubt on the notion that voter ID laws foster integrity and confidence. The Wisconsin experience demonstrates that a preoccupation with mostly phantom election fraud leads to real incidents of disenfranchisement, which undermine rather than enhance confidence in elections, particularly in minority communities. To put it bluntly, Wisconsin’s strict version of voter ID law is a cure worse than the disease. But I must follow Frank and Crawford and reject plaintiffs’ facial challenge to the law as a whole. The most pointed problem with Wisconsin’s voter ID law is that it lacks a functioning safety net for qualified electors who cannot get a voter ID with reasonable effort. The IDPP is supposed to be this safety net, but as Mrs. Smith’s story illustrates, the IDPP is pretty much a disaster. It disenfranchised about 100 qualified electors—the vast majority of whom were African American or Latino—who should have been given IDs to vote in the April 2016 primary. But the problem is deeper than that: even voters who succeed in the IDPP manage to get an ID only after surmounting severe burdens. If the petitioner lacks a birth certificate and does not have one of the usual alternatives to a birth certificate, on average, it takes five communications with the DMV after the initial application to get an ID. I conclude that the IDPP is unconstitutional and needs to be reformed or replaced. Because time is short with the fall elections approaching, I will issue an injunction targeted to the constitutional deficiencies that I identify. Judge Lynn Adelman for the U.S. District Court for the Eastern District of Wisconsin has also concluded that the IDPP is likely unconstitutional, and he has issued a preliminary injunction requiring Wisconsin to institute an affidavit procedure. This procedure would 4 allow an elector without an ID to vote by signing an affidavit stating that he or she is a qualified elector but could not get a photo ID. Judge Adelman’s injunction provides one type of safety net. But plaintiffs have not asked me to impose that solution, and I will not. The state has already issued an emergency rule under which those who are in the IDPP will get receipts valid for voting. Although that is not a complete or permanent solution, it blunts the harshest effects of the IDPP. I will also order the state to publicize that anyone who enters the IDPP will promptly get a receipt valid for voting. To address this problem over the longer term, I will order the state to reform the IDPP to meet certain standards, leaving it to the state to determine how best to cure its constitutional problems. I take this approach because it respects the state’s decision to have a strict voter ID law rather than an affidavit system. But Wisconsin may adopt a strict voter ID system only if that system has a well-functioning safety net, as both the Seventh Circuit and the Wisconsin Supreme Court have held. The heart of the opinion considers whether each of the other challenged provisions unduly burdens the right to vote, in violation of the First and Fourteenth Amendments. This analysis proceeds under what is known as the Anderson-Burdick framework, which sets out a three-step analysis. First, I determine the extent of the burden imposed by the challenged provision. Second, I evaluate the interest that the state offers to justify that burden. Third, I judge whether the interest justifies the burden. Certain of Wisconsin’s election laws fail Anderson-Burdick review. For reasons explained in the opinion, I conclude that the state may not enforce:  most of the state-imposed limitations on the time and location for inperson absentee voting (although the state may set a uniform rule disallowing in-person absentee voting on the Monday before elections);  the requirement that “dorm lists” to be used as proof of residence include citizenship information; 5  the 28-day durational residency requirement;  the prohibition on distributing absentee ballots by fax or email; and  the bar on using expired but otherwise qualifying student IDs. The purported justifications for these laws do not justify the burdens they impose. Plaintiffs also contend that the challenged laws intentionally discriminate on the basis of race and age. This is a serious charge against Wisconsin public officials. I reject most of it, applying the framework set out by the Supreme Court in Village of Arlington Heights v. Metropolitan Housing Development Corporation. But applying that same framework, I find that 2013 Wis. Act 146, restricting hours for in-person absentee voting, intentionally discriminates on the basis of race. I reach this conclusion because I am persuaded that this law was specifically targeted to curtail voting in Milwaukee without any other legitimate purpose. The legislature’s immediate goal was to achieve a partisan objective, but the means of achieving that objective was to suppress the reliably Democratic vote of Milwaukee’s African Americans. Thus, I conclude that the limits on in-person absentee voting imposed by Act 146 fail under the Fifteenth Amendment, as well as under the Anderson-Burdick analysis. In sum, Wisconsin has the authority to regulate its elections to preserve their integrity, and a voter ID requirement can be part of a well-conceived election system. But, as explained in the pages that follow, parts of Wisconsin’s election regime fail to comply with the constitutional requirement that its elections remain fair and equally open to all qualified electors. One last point: I do not intend to disrupt the August 6, 2016 election. My decision and the injunction will have no effect on that election. 6 Contents Facts .......................................................................................................................................... 8 A. The challenged provisions ............................................................................................... 9 B. Parties and procedural history ...................................................................................... 11 Analysis ................................................................................................................................... 12 A. Standing ........................................................................................................................ 13 B. Facial challenges to Wisconsin’s voter ID law .............................................................. 18 1. Facial relief because of intentional discrimination ...................................................... 22 2. Facial relief because the IDPP has failed ..................................................................... 23 C. Intentional discrimination ............................................................................................ 29 1. Race discrimination ...................................................................................................... 29 2. Age discrimination ........................................................................................................ 45 D. Partisan fencing claim ................................................................................................... 49 E. First and Fourteenth Amendment claims for undue burdens on the right to vote ....... 53 1. Limiting in-person absentee voting .............................................................................. 55 2. Requiring documentary proof of residence and eliminating corroboration............... 63 3. Changing how students can use “dorm lists” to register ............................................ 67 4. Eliminating statewide SRDs and eliminating SRDs and registration locations at high schools ............................................................................................................... 69 5. Preempting Madison’s landlord ordinance.................................................................. 72 6. Increasing the durational residency requirement ........................................................ 74 7. Establishing a zone for election observers ................................................................... 79 8. Eliminating straight-ticket voting ................................................................................ 82 9. Prohibiting clerks from sending absentee ballots by fax or email .............................. 84 10. Limiting when clerks can return absentee ballots to voters........................................ 87 11. The IDPP ....................................................................................................................... 89 12. Cumulative effect .......................................................................................................... 91 F. Voting Rights Act claims ............................................................................................... 93 1. Disparate burdens ......................................................................................................... 94 2. Caused by or linked to social and historical conditions ........................................... 105 G. Fourteenth Amendment claims for disparate treatment of voters .............................. 111 Conclusion and Remedies ..................................................................................................... 115 Order ..................................................................................................................................... 118 7 FACTS Although extensive evidence has been presented in this case, material factual disputes few and quite circumscribed. The parties sharply dispute plaintiffs’ allegations that any of the challenged laws were motivated by improper purposes, particularly intentional race and age discrimination. The parties also dispute the effect of the challenged laws on voter turnout, and whether these effects are felt more heavily by minorities and other groups of voters. But much is undisputed. The parties have stipulated to a set of background facts, most of which describe the challenged provisions and how they operate. See Dkt. 184. The court adopts these facts and recounts them below, along with other facts about Wisconsin’s election system before the challenged provisions went into effect. The court also adopts the facts found by Judge Adelman concerning the history and operation of the IDPP, which he based substantially on the evidence presented in this case. Frank v. Walker, No. 11-cv-1128, 2016 WL 3948068 (E.D. Wis. July 19, 2016). The court will incorporate the rest of its factual findings in the analysis section of this opinion. Historically, Wisconsin has had a well-respected election system, and the state has consistently had turnout rates among the highest in the country. Presidential elections were close in Wisconsin: the 2000 and 2004 elections were decided by less than one-half of one percentage point. In 2008, however, President Obama won Wisconsin by almost 14 percentage points. Two years later, Republicans took control of both houses of the state legislature, and voters elected a Republican governor. Since then, Wisconsin has implemented a series of election reforms. These laws covered almost every aspect of voting: registration, absentee voting, photo identification, and election-day mechanics. 8 A. The challenged provisions On May 25, 2011, Wisconsin enacted 2011 Wis. Act 23. That legislation made the following changes to Wisconsin election law:  It imposed a voter ID requirement.  It reduced the window of time during which municipalities could offer inperson absentee voting from a period of as much as 30 days that ended on the day before election day to a period of 12 days that ended on the Friday before election day.  It eliminated “corroboration” as a means of proving residence for the purpose of registering to vote.2  It mandated that any “dorm list” provided to a municipal clerk to be used in connection with college IDs to prove residence for the purpose of registering to vote include a certification that the students on the dorm list were United States citizens.  It increased the in-state durational residency requirement for voting for offices other than president and vice president from 10 days to 28 days before an election and required individuals who moved within Wisconsin later than 28 days before an election to vote in their previous wards or election districts.  It eliminated straight-ticket voting on official ballots.  It eliminated the authority of the Government Accountability Board (GAB) to appoint special registration deputies (SRDs) who could register voters on a statewide basis. On November 16, 2011, Wisconsin enacted 2011 Wis. Act 75, which prohibited municipal clerks from faxing or emailing absentee ballots to absentee voters other than overseas and military voters. 2 Corroboration allows a registered voter to sign a statement verifying the residence of another person, which allows that person to register to vote. 9 On April 6, 2012, Wisconsin enacted 2011 Wis. Act 227, which prohibited municipal clerks from returning an absentee ballot to an elector unless the ballot was spoiled or damaged, had an improperly completed certificate, or had no certificate. Also on April 6, 2012, Wisconsin enacted 2011 Wis. Act 240, which eliminated the requirements that SRDs be appointed at public high schools; that, in certain circumstances, SRDs be appointed at or sent to private high schools and tribal schools; and that voterregistration applications from enrolled students and members of a high school’s staff be accepted at that high school. In August 2012, the GAB directed election officials to accept electronic versions of documents that could be used to prove residence for the purpose of registering to vote. On March 20, 2013, Senate Bill 91 was introduced in the Wisconsin State Senate. This bill would have permitted municipalities to open multiple in-person absentee voting locations (under existing law, municipalities were limited to only one location). The bill failed to pass. On December 12, 2013, Wisconsin enacted 2013 Wis. Act 76. This legislation had the effect of overturning a city ordinance in Madison that required landlords to provide voter-registration forms to new tenants. On March 27, 2014, Wisconsin enacted 2013 Wis. Act 146, which reduced the window during which municipalities could offer in-person absentee voting. This law eliminated the option of offering in-person absentee voting on weekends and on weekdays before 8 a.m. or after 7 p.m. On April 2, 2014, Wisconsin enacted 2013 Wis. Act 177, which required that observation areas at polling places be placed between three and eight feet from the location 10 where voters signed in and obtained their ballots and from the location where voters registered to vote. Also on April 2, 2014, Wisconsin enacted 2013 Wis. Act 182, which required all voters, other than statutory overseas and military voters, to provide documentary proof of residence when registering to vote. Before the passage of this legislation, the requirement that a voter provide documentary proof of residence when registering to vote applied only to those who registered after the third Wednesday preceding (i.e., 20 days before) an election. B. Parties and procedural history The plaintiffs in this case include two organizations and several individuals. One Wisconsin Institute, Inc. is a nonprofit corporation with a mission “to advance progressive values, ideas, and policies through strategic research and sophisticated communications.” Dkt. 141, ¶ 4. Citizen Action of Wisconsin Education Fund, Inc. is also a nonprofit corporation focused on pursuing social and economic justice. The individual plaintiffs are Renee Gagner, Anita Johnson, Cody Nelson, Jennifer Tasse, Scott Trindl, Michael Wilder, Johnny Randle, David Walker, David Aponte, and Cassandra Silas. They all allege that the challenged provisions injure their rights to vote, register to vote, register others to vote, or vote for Democratic candidates. The initial defendants in this case were the members of the GAB and two of its officers. Plaintiffs have added and removed some defendants along the way, and the list now includes: Mark Thomsen, Ann Jacobs, Beverly Gill, Julie Glancey, Steve King, and Don Mills, the members of the Wisconsin Elections Commission; Michael Haas, the administrator of the Wisconsin Elections Commission; Mark Gottlieb, the secretary of the Wisconsin Department 11 of Transportation (DOT); and Kristina Boardman, the administrator of the DMV. Plaintiffs have sued all defendants in their official capacities. Plaintiffs filed this suit in May 2015, alleging that the challenged provisions were unconstitutional, violated the Voting Rights Act, and resulted from intentional discrimination by the Wisconsin legislature. The court granted defendants’ motion to dismiss plaintiffs’ challenge to the voter ID law, as well as some of their Equal Protection challenges to other provisions. Dkt. 66. But the court later permitted plaintiffs to partially reinstate their claims regarding the voter ID law, based on evidence that defendants produced during discovery. Dkt. 139. A few months later, the court substantially denied defendants’ motion for summary judgment, Dkt. 185, and the case proceeded to trial. ANALYSIS The court will structure its analysis as follows: First, standing. The court concludes that plaintiffs have standing to challenge each of the provisions at issue, and that the corporation plaintiffs can pursue claims under the Voting Rights Act. Second, plaintiffs’ facial challenges to Wisconsin’s voter ID law. This law has already been upheld after extensive litigation in the federal courts. The court concludes that invalidating the entire voter ID law would not be appropriate in this case. Third, plaintiffs’ claims of intentional discrimination. Plaintiffs have proven by a preponderance of the evidence that the legislature passed the provisions limiting the hours for in-person absentee voting at least partially with the intent to discriminate against voters on the basis of race. But the court concludes that the remaining provisions do not violate the 12 Fifteenth Amendment. The court also concludes that none of the challenged provisions violate the Twenty-Sixth Amendment. Fourth, plaintiffs’ “partisan fencing” claims. Although plaintiffs allege a separate claim for partisan fencing, the court concludes that their constitutional claim provides an adequate framework for analyzing these allegations. Fifth, plaintiffs’ First and Fourteenth Amendment claims for unduly burdening the right to vote. The court concludes that some, but not all, of the challenged provisions are unconstitutional because the state’s justifications for them do not outweigh the burdens that they impose. Sixth, plaintiffs’ Voting Rights Act claims. The court concludes that one of the challenged provisions violates the Voting Rights Act. Seventh, plaintiffs’ Fourteenth Amendment Equal Protection claim. The court concludes that defendants have failed to articulate a rational basis for the state’s decision to exclude expired student IDs as acceptable forms of voter ID. A. Standing The court begins with standing. At summary judgment, the court rejected defendants’ justiciability arguments, including arguments related to standing. Defendants now renew some of these arguments, contending that no plaintiff has standing to challenge the voter ID law. Defendants also contend that plaintiffs lack standing to challenge almost all of the other provisions that are at issue. For plaintiffs’ Voting Rights Act claims, defendants contend that no plaintiff qualifies as an “aggrieved person” able to pursue claims under the act. “[T]he ‘irreducible constitutional minimum’ of standing consists of three elements. The plaintiff must have (1) suffered an injury in fact, (2) that is fairly traceable to the 13 challenged conduct of the defendant, and (3) that is likely to be redressed by a favorable judicial decision.” Spokeo, Inc. v. Robins, 136 S. Ct. 1540, 1547 (2016) (citation omitted), as revised, (May 24, 2016). Defendants contend that plaintiffs have not proven the first of these elements: a cognizable injury in fact. As the parties invoking this court’s jurisdiction, plaintiffs bear the burden of establishing that they have standing. Id. But only one plaintiff needs to have standing to challenge a given provision because the complaint seeks only injunctive relief. Crawford v. Marion Cty. Election Bd., 472 F.3d 949, 951 (7th Cir. 2007), aff’d, 553 U.S. 181 (2008). Of the 10 individual plaintiffs in this case, 6 received qualifying IDs from the DMV and 4 received receipts through the IDPP. DX022; PX445. Defendants want to stop there, arguing that none of the individual plaintiffs are harmed by the voter ID law because they all currently have qualifying IDs. But there are several problems with this argument. The most obvious problem is that under the DMV’s current rules, the receipts that four of the individual plaintiffs received will expire after two automatic renewals, which means 180 days after issuance. Although these plaintiffs will be able to vote in the upcoming August and November elections, there is essentially no plan in place for them after they use their two renewals. Without a valid ID, these plaintiffs will not be able to vote. Thus, they have “suffered ‘an invasion of a legally protected interest’ that is ‘concrete and particularized’ and ‘actual or imminent, not conjectural or hypothetical.’” Spokeo, 136 S. Ct. at 1548. Even setting aside the plaintiffs who will lack acceptable IDs and be unable to vote after the November 2016 election, the voter ID law also injures the remaining individual plaintiffs. At summary judgment, the court concluded that having to present an ID at the polls was a sufficient injury for purposes of conferring Article III standing. Dkt. 185, at 10 (citing 14 Frank v. Walker, 17 F. Supp. 3d 837, 866 (E.D. Wis.), rev’d, 768 F.3d 744 (7th Cir. 2014), cert. denied, 135 S. Ct. 1551 (2015), and Common Cause/Georgia v. Billups, 554 F.3d 1340, 1351-52 (11th Cir. 2009)). The court also concluded that the plaintiffs who have IDs will have to renew them or acquire other forms of identification once their current IDs expire, which would be another injury that confers standing. Id. Defendants do not substantively engage these issues; they simply assert that “[t]his Court was wrong when it held that voters who have a qualifying ID have Article III standing to challenge the voter photo ID law.” Dkt. 206, at 13. If defendants want to preserve the issue for appeal, then they have done so. But they have not identified reasons for the court to depart from its earlier conclusion that plaintiffs have standing to challenge the voter ID law. As for the other provisions at issue, the corporation plaintiffs have standing to challenge these laws. “An organization may establish an injury to itself sufficient to support standing to challenge a statute or policy by showing that the statute or policy frustrates the organization’s goals and necessitates the expenditure of resources in ways that would not otherwise be required.” 15 James Wm. Moore et al., Moore’s Federal Practice § 101.60[1][f] (3d ed. 2015) (citing Havens Realty Corp. v. Coleman, 455 U.S. 363, 379 (1982)); see also Crawford, 472 F.3d at 951 (“[T]he new law injures the Democratic Party by compelling the party to devote resources to getting to the polls those of its supporters who would otherwise be discouraged by the new law from bothering to vote.”). To establish standing, an organization must point “to a ‘concrete and demonstrable injury to its activities,’ not ‘simply a setback to the organization’s abstract social interests.’” Spann v. Colonial Vill., Inc., 899 F.2d 24, 27 (D.C. Cir. 1990) (alterations omitted) (quoting Havens Realty Corp., 455 U.S. at 379). 15 At trial, plaintiffs adduced evidence that One Wisconsin and Citizen Action each devoted money, staff time, and other resources away from their other priorities to educate voters about the new laws. For example, Analiese Eicher, One Wisconsin’s program and development director, testified that she researched all but one of the challenged provisions. Tr. 5p, at 145:12-17.3 The purpose of this research was to allow One Wisconsin to educate its supporters, its partners, and the press. Id. at 145:18-25. Eicher also testified that had she not been researching the legislation, she would have been working on other programs or initiatives for One Wisconsin. Id. at 147:4-16. Eicher would have been advocating for other voting-related changes, such as automatic voter registration, online registration, and felony reenfranchisement. Id. at 147:18-24. On an organizational level, One Wisconsin developed a website to help voters navigate the registration process in an effort to remediate some of the confusion surrounding the challenged provisions. Id. at 148:7-9, 149:3-8. Likewise, Anita Johnson, an individual plaintiff and one of Citizen Action’s community organizers, testified that her job responsibilities have “ballooned” over the last few years as the laws have changed. Tr. 1p, at 4:16-5:1. Her presentations to community groups now take longer, she has been able to register fewer people, and she has stopped working on other issues for Citizen Action to focus exclusively on voting rights. Id. at 5:1516, 7:20-8:5, 11:7-25, 32:24-33:11. Based on this evidence, the court finds that the corporation plaintiffs are not simply redirecting their resources to litigation, which would not be an injury-in-fact that would confer standing. See N.A.A.C.P. v. City of Kyle, 626 F.3d 233, 238 (5th Cir. 2010). Instead, 3 Citations to trial transcripts are by day, session, page, and line. Thus, “Tr. 5p, at 145:1217” refers to the transcript from the fifth day of trial, afternoon session, page 145, lines 12 through 17. 16 both corporations are devoting resources away from other tasks and toward researching, or educating voters about, the challenged provisions. These expenditures are injuries that give both corporations standing to challenge the provisions at issue in this case because the corporations are counteracting what they perceive to be unlawful practices. Cf. Fla. State Conference of N.A.A.C.P. v. Browning, 522 F.3d 1153, 1166 (11th Cir. 2008). Defendants’ final justiciability challenge relates to the Voting Rights Act and whether any plaintiff qualifies as an “aggrieved person” for purposes of bringing suit pursuant to 52 U.S.C. § 10302. The court rejected this challenge at summary judgment, adopting the Eastern District of Wisconsin’s reasoning in Frank and concluding that the corporation plaintiffs could assert claims under the Voting Rights Act. Dkt. 185, at 14-15. Once again, defendants do not substantively confront this analysis. See Dkt. 206, at 15. In fact, the authority on which defendants rely—Roberts v. Wamser, 883 F.2d 617 (8th Cir. 1989)—does not actually support their assertion that corporations cannot file suit under the Voting Rights Act. Roberts involved an unsuccessful political candidate whose alleged injury was the loss of votes that he would have received but for the challenged voting practice. 883 F.2d at 621. The Eighth Circuit held “that an unsuccessful candidate attempting to challenge election results does not have standing under the Voting Rights Act.” Id. But the Eighth Circuit also noted that the candidate was not suing on behalf of others who were unable to protect their own rights, id., which is what the corporation plaintiffs are doing in this case. The court will adhere to its earlier conclusion that One Wisconsin and Citizen Action can pursue claims under the Voting Rights Act. 17 B. Facial challenges to Wisconsin’s voter ID law Wisconsin’s voter ID law has been through the federal courts before. The Seventh Circuit upheld the law in Frank v. Walker, 768 F.3d 744 (7th Cir. 2014), cert. denied, 135 S. Ct. 1551 (2015), relying on the Supreme Court’s decision in Crawford v. Marion County Election Board, 553 U.S. 181 (2008). Thus, this court will begin its consideration of the merits by addressing plaintiffs’ contention that despite the holdings in Crawford and Frank, Wisconsin’s voter ID law is facially unconstitutional and violates the Voting Rights Act. Crawford considered a facial challenge to Indiana’s voter ID law. 553 U.S. at 185. The critical holding in Crawford is that requiring a voter to show a photo ID before voting serves the important governmental interest in ensuring the integrity of elections, particularly by preventing in-person voting fraud, thereby engendering confidence in elections. Id. at 200-03. Crawford also held that securing an Indiana photo ID, which required assembling certain vital documents and going to the DMV to apply for the ID, imposed only modest burdens that were not much greater than the effort ordinarily required to register and vote. Id. at 198. Crawford upheld Indiana’s voter ID law against a facial challenge even though the burdens of the law fell somewhat more heavily on minority voters, and even though some individual voters might not be able to get a photo ID without surmounting more severe burdens. In Frank, the Seventh Circuit considered a facial challenge to Wisconsin’s voter ID law. 768 F.3d at 745. The district court had determined that there were factual distinctions between Wisconsin’s law and Indiana’s law: most significantly, that there were many more voters who did not have a qualifying photo ID in Wisconsin, and that those voters tended to be minorities. The Seventh Circuit expressed skepticism about the evidence of how many voters lacked ID, but concluded that, in any case, those distinctions were not material to the 18 facial challenge. The Seventh Circuit held that Wisconsin’s voter ID law was not materially different from the Indiana law at issue in Crawford, and that under Crawford, Wisconsin’s voter ID law was facially constitutional. Id. It is hard to deny that a state and its citizens have a truly compelling interest in maintaining election integrity. As the evidence in this case proved once again, voter fraud is rare but not non-existent. The court credits the evidence of plaintiffs’ expert on the subject, Dr. Lorraine C. Minnite, who testified and filed two expert reports. PX039; PX044. But the more compelling evidence comes from Milwaukee County, the one county in the state that has tried to systematically discover and track violations of election law. The county has an assistant district attorney devoted full-time to the job, Bruce Landgraf. Based on Landgraf’s testimony, and on other evidence discussed below, the court finds that impersonation fraud— the type of fraud that voter ID is designed to prevent—is extremely rare. In most elections there are a very few incidents in which impersonation fraud cannot be ruled out. But as Crawford and Frank held, despite rarity with which election fraud occurs, it is nevertheless reasonable for states to take steps to prevent it. Any system that requires voters to get a credential will necessarily impose a burden on them. But if the burden is a modest one, and if the credential meaningfully fosters integrity, then the constitution is satisfied. Under Crawford and Frank, collecting the necessary records and making a trip to the DMV to get an ID is a modest burden in light of the state interest that it serves. Those cases probably reflected an unduly rosy view of DMV field offices, but the evidence in this case confirms, yet again, that the vast majority of Wisconsin citizens already have the necessary ID. And most citizens who do not have an ID can get one with relative ease. 19 This court is, of course, bound to follow Crawford and Frank, which defendants contend doom plaintiffs’ facial challenge to Wisconsin’s voter ID law. Defendants are correct. But Crawford and Frank deserve reappraisal. The court is skeptical that voter ID laws engender confidence in elections, which is one of the important governmental purposes that courts have used to sustain the constitutionality of those laws. The evidence in this case showed that portions of Wisconsin’s population, especially those who live in minority communities, perceive voter ID laws as a means of suppressing voters. This means that they undermine rather than enhance confidence in our electoral system. Good national research suggests that voter ID laws suppress turnout, and that they have a small, but demonstrable, disparate effect on minority groups. See PX072. At trial, testimony of African American community leaders confirmed that voter ID laws engender acute resentment in minority communities. See, e.g., Tr. 1p, at 131:21-24. And some of the Wisconsin legislators who supported voter ID laws believed that they would have partisan effects. Their willingness to publically tout the partisan impact of those laws deepens the resentment and undermines belief in electoral fairness. Underlying the philosophical debate is a fundamentally factual question: do voter ID laws protect the integrity of elections? According to the Frank court, Crawford definitively answered this question. 768 F.3d at 750 (“[W]hether a photo ID requirement promotes public confidence in the electoral system is a ‘legislative fact’—a proposition about the state of the world, as opposed to a proposition about these litigants or about a single state.”). The primary integrity-based justification offered for voter ID laws is that they prevent voter fraud. But that seems to be a dubious proposition. A voter ID requirement addresses only certain types of election malfeasance; specifically, impersonation fraud, by which one person poses as 20 another and votes under his or her name. This happens from time to time by accident, when a voter signs the poll book on the wrong line. That produces some frustration for voters and poll workers, but it does not represent a fundamental threat to the integrity of elections because it does not happen that often and because everyone ultimately gets to vote. The real fear is multiple voting: that a committed but unethical partisan could cast many votes for his or her candidate under different names. Yet there is utterly no evidence that this is a systematic problem, or even a common occurrence in Wisconsin or anywhere in the United States. PX039, at 2, 35. True, it is not unheard of: in one well-known case, a Milwaukee man was so committed to Governor Walker’s re-election that he voted 14 times. Tr. 8a, at 184:3-24. He was charged with and convicted of voter fraud (even without the benefits of the voter ID law). Proponents of voter ID would say that there could be other incidents of voter fraud that have gone undetected. But there is no evidence to support that hypothesis. As many have pointed out, multiple voting is not a very effective way of influencing an election, and few people would risk the penalties to do so. The bottom line is that impersonation fraud is a truly isolated phenomenon that has not posed a significant threat to the integrity of Wisconsin’s elections. The same cannot be said for Wisconsin’s voter ID law, which has so far been implemented in a rigorously strict form: the only way to vote is to secure a state-approved ID. As part of Act 23, Wisconsin enacted a statute allowing citizens to receive free IDs to vote. But it was not until the eve of trial in this case that the state started paying for the underlying documents (e.g., birth certificates) that citizens needed to submit to obtain these free IDs. Even now, citizens who lack vital records can obtain free IDs only after navigating the complicated IDPP. Wisconsin’s strict implementation of its voter ID law has 21 disenfranchised more citizens than have ever been shown to have committed impersonation fraud. In theory, the well-designed and easy-to-use registration and voting system imagined in Crawford and Frank facilitates public confidence without eroding participation in elections. But in practice, Wisconsin’s system bears little resemblance to that ideal. So where does that leave plaintiffs’ facial challenge to the voter ID law? Plaintiffs contend that two aspects of the factual record of this case distinguish it from Crawford and Frank, paving the way to a fresh facial challenge. 1. Facial relief because of intentional discrimination First, plaintiffs assert that Wisconsin’s voter ID law was motivated, at least in part, by racial animus. This is a serious allegation against the public officials of Wisconsin, but the court cannot easily dismiss it here. There is manifest racial disparity in the operation of the IDPP: of the 61 actual denials that the DMV had issued as of April 2016, 85 percent were to African Americans or Latinos. PX475. And government witnesses concede that 60 of these denials were issued to qualified electors entitled to vote, but who could not meet the IDPP’s criteria for a state-issued ID. See Tr. 6, at 75:24-76:17 (DMV administrator); Tr. 8p, at 191:2-5 (investigations unit employee). The legislative history suggests that some of the provisions challenged in this case were specifically intended to curtail voting in Milwaukee, where 40 percent of the population is African American and 17.3 percent is Latino (approximately two-thirds of the state’s minority population). Both sides agree that if the court finds that the Wisconsin legislature enacted a voter ID law for the at least partially with the intent to discriminate on the basis of race, then the law is constitutionally unsound and 22 cannot stand. The court will address this issue below, in discussing the intentional discrimination claims that plaintiffs have alleged in this case. 2. Facial relief because the IDPP has failed The second factual distinction concerns the IDPP, which plaintiffs contend imposes severe and discriminatory burdens on some qualified Wisconsin electors. The IDPP was the subject of a great deal of testimony at trial, and it has become a dominant issue in this case. Plaintiffs contend that the IDPP demonstrates Wisconsin’s intentional race discrimination, is unconstitutional under the Anderson-Burdick framework, and violates the Voting Rights Act.4 And because this constitutionally required safety net is not working, plaintiffs argue that the court must strike down the entire voter ID law. The context for, and history of, Wisconsin’s effort to implement the IDPP began with Act 23, passed in 2011. Besides establishing voter ID, this legislation created Wis. Stat. § 343.50(5)(a)3., which provided that a voter could get a Wisconsin ID from the DMV for free, if the voter requested it for voting. But voters who did not have their birth certificates had to get copies, which typically required paying a fee to a government agency. Thus, getting a free ID was not really free. Many thought that the fees that voters had to pay for copies of their vital records were tantamount to an unconstitutional poll tax. Indeed, that was the conclusion that the Wisconsin Supreme Court reached in Milwaukee Branch of NAACP v. Walker, which relied on Crawford to uphold Wisconsin’s voter ID law against a facial challenge. 2014 WI 98, ¶ 7, 357 Wis. 2d 469, 851 N.W.2d 262, reconsideration dismissed, 856 N.W.2d 177 (2014). The state supreme court applied a savings construction to the Wisconsin Administrative Code to 4 The court will analyze the IDPP under these legal theories later in this opinion. 23 provide that the required vital documents were “unavailable” to a prospective voter if he or she would have to pay a fee to get them. Id. ¶¶ 66-71. Thus, a person who had to pay to get a birth certificate could use the DMV’s special petition process in Wis. Admin. Code DOT § 102.15 (i.e., the IDPP) to ask for a free ID on the grounds that a birth certificate was unavailable. As the Seventh Circuit recognized in Frank, the availability of a truly free ID provided a necessary safety net that preserved the constitutionality of Wisconsin’s voter ID law. 768 F.3d at 747. But since then, effectuating the savings construction to provide free photo IDs to voters who lacked the requisite vital records has proven to be difficult for the DMV, to say the least. For purposes of this opinion, the court does not need to retrace every detail of DOT’s response to NAACP v. Walker; plaintiffs have set out the timeline in a chart appended to their brief. Dkt. 207, at 253-57. In summary, the DOT instituted an emergency rule on September 11, 2014 (the day before the appellate argument in Frank). PX456. The emergency rule changed the definition of “unavailable,” following the Wisconsin Supreme Court’s direction, and it reorganized the IDPP into a new subsection of Wisconsin’s Administrative Code, DOT § 102.15(5m). The emergency rule also created a procedure that, in essence, required the DMV to track down the birth record of any person who requested a free voter ID, if the person did not have a copy of their birth record. The procedure was complicated because the process required interaction between various divisions of the DMV, the Wisconsin Department of Health Services, and agencies of other states. PX472. The main task of investigating and evaluating petitions fell to the DMV’s Compliance and Fraud Unit (CAFU), which, as its name implies, has staff members whose normal duties are to investigate allegations of fraud. 24 Many people successfully navigated the IDPP. Out of 1,389 petitions for free IDs, the DVM issued IDs to 1,132 petitioners. Of the petitioners who applied, 487 had to go through “adjudication,” which included a full investigation by CAFU5 and a final decision from Jim Miller, the head of the DMV’s Bureau of Field Services (a different unit from CAFU). 230 of the petitioners who went through adjudication received IDs; 257 petitioners did not. DMV records indicate that 98 of the petitioners who did not receive IDs after adjudication cancelled their petitions.6 The petitioners in suspended or denied status were the ones who faced serious roadblocks in the IDPP: their birth records did not exist, or those records did not perfectly match their names or other aspects of their identities, such as Social Security records. The problems arose because the DMV evaluated IDPP petitions for voting IDs by using the same identification standards that it applied to applications for Wisconsin driver licenses and standard IDs. To acquire any one of these products from the DMV, a person must prove both their identity and their legal presence in the United States. Thus, the DMV refused to issue IDs to IDPP petitioners until CAFU could confirm their identities with a match to a 5 Full investigation by CAFU commonly involved acquiring a CLEAR background report. These reports contained a substantial amount of deeply personal information, including any criminal records, judgments and liens, residence history, home and vehicle ownership history, and a list of possible relatives and associates. The DMV witnesses testified that the DMV never used CLEAR reports to the disadvantage of petitioners. But even assuming that CLEAR reports were acquired only to connect petitioners to vital records, the court finds that having DMV personnel acquire and review a compilation of personal information imposes a substantial burden on the right to vote. 6 The DMV’s code for “customer initiated cancel” covers a wide range of results. For example, petitions received this code when the petitioner died while the petition was pending. Petitions also received this code if a petitioner simply gave up or if he or she found a birth certificate and applied for a standard state-issued ID. 25 valid birth record, or to some equivalently secure alternative. Some petitioners simply could not meet the DMV’s standard of proof, and so they could not obtain free IDs. The lack of a valid birth record correlated strikingly, yet predictably, with minority status. The evidence at trial demonstrated that Puerto Rico, Cook County, Illinois, and states with a history of de jure segregation have systematic deficiencies in their vital records systems. Voters born in those places were commonly unable to confirm their identities under the DMV’s standards. For example, many African American residents in Wisconsin were born in Cook County or in southern states. PX479. And many of the state’s Latino residents were born in Puerto Rico. Id. As of April 2016, more than half of the petitioners who had entered the IDPP were born in Illinois, Mississippi, or a southern state that had a history of de jure segregation. PX478. In June 2015, the DMV begin issuing denials to IDPP petitioners. By the time of trial in this case, the DMV had issued 61 denials, 53 of which were to minority petitioners.7 Again, with one exception, the DMV had no reason to doubt that those who were denied a photo ID were Wisconsin residents, United States citizens, at least 18 years of age, and qualified to vote. Tr. 6, at 75:24-76:17. The sole exception was a Latina woman who mistakenly believed that she had been naturalized. Since the state first implemented the IDPP, another related problem has prevented petitioners from successfully navigating the process. Until recently, the state had not appropriated any funds to pay for petitioners’ vital records. Although no petitioner was asked to pay for any vital record, the state did not acquire any vital record for which a fee was 7 Nine of the petitioners who received denial letters were able to track down vital records on their own and receive free IDs without using the IDPP. See Dkt. 207, at 69 (discussing examples). The DMV re-coded these denials to “customer-initiated cancellations.” 26 required. The result was that some petitioners fell into limbo: the DMV did not deny their petitions, but the petitioners could not confirm their identities. These petitioners ended up in “suspend” status, with the DMV essentially waiting either for the petitioner to turn up new records, or for enough time to pass that the DMV could officially deny the petition. On March 7, 2016, DMV officials and state legal counsel met to discuss the state’s failure to pay for vital records. At some point after the meeting, the DMV received funds, and during the second week of trial in this case, the DMV made its first payment to acquire a vital record for a petitioner. Tr. 7p, at 111:2-17. On May 10, 2016, a week before the trial in this case began, the governor approved another emergency rule modifying the IDPP. PX452. The new rule acknowledged that emergency rulemaking was required to ensure that qualified electors could get a photo ID with reasonable effort in time for the next elections: This emergency rulemaking [was] also necessary to preserve the integrity of the verification process utilized by the Department in issuing an identification card while still preserving the public welfare by ensuring that qualified applicants who may not be able to obtain acceptable photographic identification for voting purposes with reasonable effort will be able to obtain photographic identification before the next scheduled elections. PX453, at 14. The rule ameliorated some of the deficiencies of the IDPP: it established procedures and standards for evaluating petitions; it provided a means to surmount common impediments such as minor mismatches between a birth record and other aspects of a petitioner’s identity; and it established “more likely than not” as the standard for evaluating evidence of identity, birthdate, and citizenship.8 Perhaps most important, the emergency rule 8 At trial, DMV witnesses testified that the new emergency rule codified current practice. Tr. 8p, at 190:7-193:7. This testimony was not credible. The testimony of CAFU employees showed that petitioners were held to a much higher standard than “more likely than not.” 27 required the DMV to issue petitioners temporary identification card receipts that were valid for voting purposes while their petitions were pending. Defendants contend that the latest emergency rule fixes the problems with the IDPP, and that because all petitioners still in the process have a receipt valid for voting, the dispute over the IDPP is moot. The court disagrees for two reasons. First, the receipts issued under the emergency rule are not permanent. Those who hold them will be able to vote only so long as the receipts are renewed. But qualified electors are entitled to vote as a matter of constitutional right, not merely by the grace of the executive branch of the state government. The state has promised to renew the receipts for 180 days so that they will be good through the November 2016 election. But the state has been utterly silent on what happens after that. As things stand now, after these receipts expire, petitioners will once again find themselves in IDPP limbo. Thus, at best, the emergency rule gives the state time to devise a new solution (but the court has not seen any evidence to suggest that the state is actually working on a solution). Second, even under the emergency rule, petitioners will have to convince the DMV to exercise its discretion to issue them IDs. Although the emergency rule guides that discretion and specifies that the applicable standard of proof is “more likely than not,” the process is still far more arduous than collecting documents and making a trip to the DMV, as envisioned in Crawford and Frank. Being investigated by CAFU, even under the newest iteration of Wisconsin’s emergency rule, still makes it unnecessarily difficult to obtain an ID. The court finds that IDPP petitions were decided by a standard that was at least as rigorous as “clear and convincing proof.” 28 For now, suffice it to say that the court agrees that the IDPP is a wretched failure: it has disenfranchised a number of citizens who are unquestionably qualified to vote, and these disenfranchised citizens are overwhelmingly African American and Latino. The IDPP violates the constitutional rights of those who must use it, and so Wisconsin must therefore replace or substantially reform the process. But that does not mean that the voter ID law is unconstitutional in all of its applications. Because a targeted remedy can cure the constitutional flaws of the IDPP (and thus, the entire voter ID law), facial relief is not necessary or appropriate. Crawford and Frank effectively foreclose invalidating Wisconsin’s voter ID law outright. Based on the evidence presented at trial, the court has some misgivings about whether the law actually promotes confidence and integrity. But precedent is precedent, and so the court will deny plaintiffs’ request to invalidate the entire voter ID regime. C. Intentional discrimination Plaintiffs assert claims under the Fifteenth and Twenty-Sixth Amendments, alleging intentional discrimination on the basis of race and on the basis of age. The legal standards for evaluating these claims are substantially identical, and most of the pertinent evidence for each claim is the same. With the exception of Wisconsin’s restriction on the number of hours that municipal clerks can offer in-person absentee voting, the court concludes that plaintiffs have failed to prove their claims of intentional discrimination. 1. Race discrimination Plaintiffs contend that the Wisconsin legislature passed many of the challenged provisions in violation of the Fifteenth Amendment. To succeed on these claims, plaintiffs must demonstrate that the legislature intentionally discriminated against voters because of 29 their race. Rogers v. Lodge, 458 U.S. 613, 617 (1982); Village of Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252, 265 (1977). Discriminatory animus does not need to be the only reason for Wisconsin’s new laws, or even the primary reason,but “official action will not be held unconstitutional solely because it results in a racially disproportionate impact.” Arlington Heights, 429 U.S. at 264-65. Nor do plaintiffs have to prove discriminatory intent with direct evidence of racial animus. Rogers, 458 U.S. at 618. Whether a law is motivated by racial discrimination is a difficult factual determination, guided by sparse precedent. Arlington Heights provides the essential template: “Determining whether invidious discriminatory purpose was a motivating factor demands a sensitive inquiry into such circumstantial and direct evidence of intent as may be available.” 429 U.S. at 266. The starting point of the analysis is whether the law has had a disparate impact. But unless there is a startling pattern, inexplicable on grounds other than race, impact alone is not determinative. In that case, other evidence must support a finding of discrimination. This evidence can include the historical background and context of the law and the legislative history, especially any contemporaneous statements by the decisionmaking body. See id. at 266-68. Before turning to the Arlington Heights analysis, the court considers defendants’ evidentiary objection to one of plaintiffs’ experts, historian Allan Lichtman, PhD. At trial, Dr. Lichtman testified that several of the challenged provisions were motivated by intentional race discrimination. See Tr. 6, at 237:5-18. Defendants contend that Dr. Lichtman’s testimony invaded the province of the court by offering an opinion on an ultimate issue in the case, and that it was therefore not a proper topic for expert analysis. The court agrees. Dr. Lichtman provided some useful factual background to the legislation at issue— 30 background that defendants did not dispute—but the court will not otherwise adopt his analysis or opinions about the specific issue of the legislature’s intent in passing the challenged provisions. With these considerations in mind, the court turns to the merits of plaintiffs’ intentional race discrimination claim. The court will analyze this claim first in the context of Wisconsin’s voter ID law, then in the context of the IDPP, and finally in the context of the other challenged provisions. a. The voter ID law To analyze whether Wisconsin’s voter ID law violates the Fifteenth Amendment, the court begins by summarizing the disparate impact that the law has had on racial minorities. The question of how many people in Wisconsin have a driver license or a Wisconsin ID has proved to be surprisingly hard to answer. The district court in Frank estimated that about 300,000, about 9 percent of the state’s registered voters, lacked a valid photo ID. 17 F. Supp. 3d at 854. The Seventh Circuit doubted this, partly because the district court in Crawford estimated that only 43,000 lacked ID in Indiana, and partly because it just seems implausible that 9 percent of the adult population could get by without a photo ID. 768 F.3d at 748. To answer this question, both sides’ experts matched the statewide voter registration database to the DMV database. Both sides recognize that the databases are not readily matched, which makes errors likely. After identifying and correcting for errors, plaintiffs’ expert, Kenneth Mayer, PhD, estimated that 8.4 percent of registered voters lack a Wisconsin ID. Defendants’ expert, M.V. Hood III, PhD, put the estimate at only 4.54 percent. The primary difference between the two experts is that Dr. Hood had the help of a DMV programmer, Fred Eckhardt, who was able to match an additional 112,817 registered 31 voters to valid Wisconsin IDs. Tr. 4p, at 201:17-202:1. The court finds that Eckhardt’s work was reliable, and that Dr. Hood’s estimate is therefore the more credible one as to the number of registered voters without ID. Unfortunately, Dr. Hood did not break those numbers down by race. Dr. Mayer did, PX038, at 19 (Table 3), and he shows that African Americans and Latinos are more likely to lack ID. But his starting point uses the inflated 8.4 percent of voters without ID. With some of its own arithmetic to reconcile Dr. Mayer’s proportions to Dr. Hood’s base,9 the court finds that approximately 4.5 percent of white voters lack ID; 5.3 percent of African American voters lack ID; and 6.0 percent of Latino voters lack ID. The court notes that these numbers say nothing about what proportions of voters lack the documentation that would allow them to get a qualifying ID if they sought one. Dr. Hood’s evidence shows that African Americans and Latinos make up a disproportionate share of those seeking free IDs for voting. African Americans accounted for 35.6 percent of free IDs, whereas they make up only 5.6 percent of the citizen voting age population. Latinos accounted for 8.3 percent of the free IDs, against only 3.3 percent of the citizen voting age population. These numbers show very pronounced racial differences among those who seek IDs. This, in turn, strongly suggests that a greatly disproportionate share of African Americans and Latinos will have to go to the trouble of acquiring a qualifying ID to vote. But most of those who seek free IDs are probably voters who have the documents necessary to get a qualifying ID. Frank recognizes that this disparity could well have a corresponding disparate effect on turnout because any procedural requirement will dissuade 9 The court also assumes that the errors corrected by Eckhardt are distributed evenly across racial groups. Nothing in Eckhardt’s description of the errors that he found suggested that they would correlate with race. 32 some voters. But under Frank, the burden of going to the DMV to get a free ID is not constitutionally significant because it is a modest burden no greater than the ordinary burdens involved in voting. Still, the evidence here shows that patterns of ID possession are racially disparate, and that is likely to have a racially disparate effect on turnout. And some proportion of those seeking IDs will lack the usual documentation and have to enter the IDPP. Those individuals, too, tend to be minorities: 67.9 percent of those who entered the IDPP were minorities. PX474. The bottom line is that the evidence suggests that the vast majority of Wisconsin voters have a qualifying ID or could get one. But both ID possession and the lack of qualifying documentation correlate strongly with race. Next, the court considers the historical background of the voter ID law. As plaintiffs showed, before 2011, Wisconsin had an exemplary election system that produced high levels of voter participation without significant irregularities. See PX036, at 23 (Lichtman report discussing studies from the Pew Charitable Trusts ranking Wisconsin second best in the nation in electoral performance in 2008 and fourth best in 2010). The court will not go so far as to say that Wisconsin could not have improved its elections. But there was no evidence that Wisconsin elections actually suffered from identifiable problems, despite unsubstantiated allegations of fraud in the 2004 presidential election. Plaintiffs contend that demographic shifts in Wisconsin made the minority vote critical to the outcome of elections. For example, from 2010 to 2014, the white voting age population in Wisconsin declined by 1.3 percent, while the African American population increased by 3.5 percent, and the Latino population increased by 8.7 percent. Id. at 16-17. Voting in Wisconsin is sharply polarized by race: in statewide elections over the last decade, 33 90 percent of African Americans and 63 percent of Latinos voted for Democratic candidates. Because Wisconsin is a closely divided swing state, marginal differences in turnout can be decisive in close elections. Plaintiffs contend that demographic and political considerations combined to give Wisconsin Republicans a motive to discriminate against minorities in voting laws. The Wisconsin political environment changed dramatically in 2010: Republican Scott Walker was elected governor, and Republicans won control of both houses of the legislature. Although the recall elections in summer 2012 briefly shifted control of the state senate to Democrats, Republicans regained control of the chamber a few months later. The legislature and the governorship have been in Republican control since then. Plaintiffs contend that sustained one-party control over the legislature and governorship gave Republicans the opportunity to pass discriminatory election legislation. Plaintiffs concede that there were no procedural irregularities in how Wisconsin’s voter ID law, or any of the other challenged provisions, were passed. “Given unified Republican control of the legislature and governorship . . . Republicans did not have to violate procedural rules to enact many of the limitations on voting” that are at issue. Id. at 48. Nevertheless, plaintiffs contend that the bills were rushed through the legislature, depriving the GAB of time to review them, and providing inadequate time for public input. See PX084. This dovetails with plaintiffs’ contention that there were substantive irregularities with the laws, by which plaintiffs mean that the laws were not well justified or consistent. Defendants are correct that the legislature had no obligation to provide any rationale to support a validly enacted law. But plaintiffs have a point: the challenged laws were passed by a process that allowed limited public input and little actual debate. The legislative history 34 demonstrates that Democrats and members of the public voiced concerns about the discriminatory impact of the laws, and that those concerns largely went unrebutted. Thus, the court has little information about what actually prompted these bills and the reasons why the legislature enacted them into law. Most of them were passed with only summary statements of legislative purpose, typically invoking only generic concerns for election integrity or consistency. See, e.g., PX058; PX216. Plaintiffs would fill the gap in the official legislative record with extra-legislative comments by Republican legislators and staffers, which plaintiffs contend strongly indicate discriminatory intent. The court will not recapitulate all such statements in the record, but plaintiffs have identified a few as particularly telling. First, plaintiffs cite to a recent comment by former state senator Glenn Grothman (now a U.S. representative) that he thought that Wisconsin’s voter ID law would help Republicans in the 2016 presidential election. PX068. Second, plaintiffs cite to Grothman’s statements on the floor of the senate in 2014 concerning the need to limit the hours for in-person absentee voting in Milwaukee. PX022. Third, plaintiffs cite to statements by former state senator Dale Schulz and by his staffer Todd Albaugh. During a radio interview, Schultz indicated that the Republican leadership of the legislature passed the voter ID law for partisan purposes, not out of any legitimate concern for the integrity of Wisconsin elections. PX067. Albaugh testified that at the last meeting of the Republican caucus before the vote on Act 23, the Republican leadership insisted that Republicans get in line to support the bill because it was important to future Republican electoral success. See Tr. 1a, at 84:1-24.10 10 At trial, defendants disputed Albaugh’s interpretation and evaluation of the meeting, and they also objected to his testimony on hearsay grounds. The court overrules the hearsay objection because Lazich’s out-of-court statements were not offered for their truth. The point 35 The parties have also stipulated to the admissibility of notes and correspondence from the files of various Republican legislators. See Dkt. 184, at 3-4. Among other things, this evidence includes senator Alberta Darling’s expressed opinion that had it been in effect, the voter ID law would have made a difference in the November 2012 election, id. at 4, which like Grothman’s more recent statement, shows that legislators believed that Act 23 would have a partisan impact on elections. The court may consider these statements under Arlington Heights. But ample authority counsels skepticism, and the court will not simplistically assign discriminatory intent to the legislature based on the comments of individual legislators. See Veasey v. Abbott, No. 1441127, 2016 WL 3923868, at *9 (5th Cir. July 20, 2016) (“While probative in theory, even those (after-the-fact) stray statements made by a few individual legislators voting for SB 14 may not be the best indicia of the Texas Legislature’s intent.”). The comments that plaintiffs have identified paint a consistent picture that resonates with the rest of the record, particularly the lack of a verified problem with voter fraud, and the increasingly partisan divisions in support for the law. The conclusion is hard to resist: the Republican leadership believed that voter ID would help the prospects of Republicans in future elections. (And for that matter, Democrats apparently thought that, too.) As for other context surrounding Wisconsin’s voter ID law, the court notes that Act 23 was the first in a series of election reforms that the Republican-controlled legislature passed between 2011 and 2014. None of these laws made registration or voting easier for was not that the voter ID law would actually help Republicans in future elections. The point was that Lazich thought they would, and that was part of her motive for encouraging support for the voter ID law. Defendants offered no evidence to dispute the accuracy of Albaugh’s recounting of what was said at the meeting. 36 anyone, but they had only minimal effect on less transient, wealthier voters. For reasons explained more fully below, the stated rationales for many provisions of Act 23, and for the election laws that followed it, were meager. Accordingly, in light of the record of the case as a whole, the conclusion is nearly inescapable: the election laws passed between 2011 and 2014 were motivated in large part by the Republican majority’s partisan interests. Against this background, the court turns to the more difficult question of whether Act 23 was motivated by racial animus. For the following reasons, the court finds that it was not. First, the legislature passed the voter ID bill in 2011, three years after the Supreme Court upheld a facial challenge to a similar voter ID law in Crawford. The Court had held that voter ID laws served a legitimate government interest in election integrity, and that they did not have an unduly disparate impact on racial minorities. Legislators would have been entitled to embrace the rationale that the Supreme Court endorsed, even if other legislators or members of the public contended that the law would have a disparate impact on minorities. Second, voter ID bills have a long history in Wisconsin and in the United States, and that history does not suggest that such laws are inherently motivated by racial animus. In 2005, the Commission on Federal Election Reform, co-chaired by Jimmy Carter and James Baker III, identified a voter ID system with photo ID as one of five pillars of a reformed U.S. election system. Commission on Federal Election Reform, Building Confidence in U.S. Elections (September 2005), http://www.eac.gov/assets/1/AssetManager/Exhibit%20M.PDF. That same year, the Wisconsin legislature passed a photo ID bill that was ultimately vetoed by Governor Doyle, a Democrat. Although Democrats tended to oppose that bill, it garnered significant 37 bipartisan support. This history shows that legislators and politicians with no motive to discriminate against minorities have nevertheless supported voter ID laws. Third, even though there is scant evidence of actual voter fraud in Wisconsin, the concern for election integrity provides a valid, non-discriminatory reason for supporting a voter ID law. To be sure, there is a legitimate countervailing concern that voter ID requirements impede access to the polls. But the existence of a robust, non-discriminatory rationale in favor of voter ID makes it hard to draw the inference that support for voter ID must be racially motivated.11 Plaintiffs nevertheless contend that the strict version of voter ID enacted in 2011 suggests a discriminatory motive. But by then, the potential for a voter ID requirement to have a racially disparate impact had long been recognized. See, e.g., id. at 20 (“The introduction of voter ID requirements has raised concerns that they may present a barrier to voting, particularly by traditionally marginalized groups, such as the poor and minorities, some of whom lack a government issued photo ID.”) Democrats, private citizens, and the GAB repeatedly raised these types of concerns to the legislature. See, e.g., PX014; PX084; PX263; PX299.The legislature passed the voter ID bill anyway, and the governor signed it. Plaintiffs contend that the legislature’s apparent willful blindness to Act 23’s disparate effects is strong evidence of discrimination. But the legislature did not entirely ignore these 11 Dr. Lichtman points out that in 2015, during consideration of a bill to require photo IDs for the Food Share program, the Wisconsin Assembly rejected an amendment that would have allowed Food Share IDs to be used for voting. PX036, at 36-37. According to Dr. Lichtman, if the legislature were sincerely interested in election integrity, it would accept Food Share IDs for voting because they are every bit as secure as Wisconsin IDs. The refusal to accept Food Share IDs is, therefore, evidence of discriminatory intent. The argument would be persuasive, if it were contemporaneous with Act 23, the voter ID law. The force of the argument is also blunted because the Food Share ID bill has not been enacted. 38 concerns. Act 23 created Wis. Stat. § 343.50(5)(a)3., which required the DMV to provide a free ID to any citizen over the age of 18 who requested one for voting. Since the introduction of the IDPP in 2014, the profound difficulty of providing traditional DMV-issued IDs to some voters has become apparent, and the state has been painfully reluctant to address these problems. But in 2011, to the legislature that passed Act 23, the free ID seemed like a reasonable response to the concerns that opponents raised. C.f. Building Confidence in U.S. Elections, at 20 (“Part of these concerns are addressed by assuring that government-issued photo identification is available without expense to any citizen.”). In sum, the court concludes that plaintiffs have not proven by a preponderance of the evidence that the voter ID provision of Act 23 was motivated, even in part, by racial animus. Wisconsin’s voter ID law therefore does not violate the Fifteenth Amendment. b. The IDPP The racial imbalances among IDPP petitioners, and among the results of the process, are striking. Minorities make up only 11 percent of Wisconsin’s citizen voting age population, but they make up 55 percent of the voters who have received free IDs since Act 23 was passed. DX265. As of April 2016, two-thirds of those who entered the process were minorities; African Americans alone represented 55.9 percent of IDPP petitioners. PX474. Worse yet, African Americans and Latinos represented 85 percent (52 out of 61) of all IDPP denials. PX475. Plaintiffs contend that these numbers present the kind of striking pattern that is inexplicable as anything but intentional discrimination. They argue that the court should find the IDPP to be unconstitutional on that basis alone, relying on decisions such as Gomillion v. 39 Lightfoot, 364 U.S. 339 (1960) (allegations of extreme gerrymandering, if proven, would be tantamount to a “mathematical demonstration” of discrimination). The court is not persuaded that statistics about the petitioners who have used the IDPP, or been denied free IDs, compel a finding of intentional race discrimination. And the reasoning is simple: the free ID procedure and the IDPP were designed to blunt the potential for disenfranchisement that might arise from Wisconsin’s voter ID law. The potential for disenfranchisement, as all recognized, fell more heavily on minorities. Thus, it is no surprise that those who sought free IDs, or who entered the IDPP because they lacked vital records, were predominantly minorities. It is also no surprise that minorities foundered at high rates in a process that required documentary proof of identity, birthdate, and citizenship. Make no mistake: the IDPP as it currently exists has failed to fulfill its constitutional purpose. But plaintiffs have not shown that it is the result of intentional race discrimination. As plaintiffs’ counsel repeatedly reiterated to the DMV witnesses, plaintiffs do not allege that DMV employees intended to discriminate against anyone. And as the court observed during trial, some CAFU employees undertook nearly heroic efforts to track down documents to prove petitioners’ identities and birthdates. The court finds that DMV employees, especially CAFU employees, undertook their duties in good faith, trying as best they could under the governing regulations to get IDs into the hands of as many petitioners as possible. Another reason why the court cannot find that the legislature intentionally discriminated on the basis of race is that the legislature did not design or implement the IDPP. The fault lies with the executive branch, which let the IDPP grind on until plaintiffs in this litigation exposed its many flaws. But plaintiffs have not shown that anyone in the executive branch knew that the IDPP was disenfranchising voters and ignored the problem. 40 The flaws would not have been hard to find, and Wisconsin should have done better. But based on the evidence presented at trial, the court cannot find that members of the executive branch acted with racial animus in creating or implementing the IDPP. c. Other challenged provisions The court now turns to the other provisions that plaintiffs challenge under the Fifteenth Amendment. Setting aside the provisions relating to in-person absentee voting, plaintiffs contend that the legislature enacted the following regulations, at least in part, with the intent to discriminate against African Americans and Latinos: (1) eliminating corroboration; (2) requiring documentary proof-of-residence; (3) eliminating statewide SRDs; (4) increasing the durational residency requirement; (5) changing the location for election observers; and (6) eliminating straight-ticket voting. Plaintiffs contend that each of these changes in Wisconsin’s voting laws particularly disadvantage minorities, who tend to be poorer, less educated, and more transient. But disparate impact alone is not enough to show intentional discrimination. Arlington Heights, 429 U.S. at 264-65. These regulations are all facially neutral, and the extra burdens that they impose would fall on anyone who is poorer, less educated, or more transient, regardless of race. As explained in other parts of this opinion, some of these regulations are not justified by significant government interests, which puts their legitimacy under Anderson-Burdick in doubt. But plaintiffs give the court no reason to find that any of these regulations were targeted at minority voters or that the legislature was racially motivated in passing any of them. Accordingly, the court concludes that plaintiffs have not shown by a preponderance of the evidence that any of these changes in Wisconsin’s voting laws were motivated, even in part, by racial animus. 41 As for the one-location rule, plaintiffs proved that forcing all municipalities to offer only one location for in-person absentee voting imposed greater burdens on voters in large municipalities like Milwaukee than it did on voters in smaller towns. And because Milwaukee has a predominantly minority population, the one-location rule was all but guaranteed to have a disparate impact. But this provision has been in effect since 2005, long before the legislature enacted the restrictions to the hours for in-person absentee voting. See Wis. Stat. § 6.855(1). Thus, the legislative history and other contextual evidence discussed above does not bear on the issue of whether the legislature passed the one-location rule with the intent to discriminate. Indeed, plaintiffs have not offered any evidence addressing the legislature’s intent in enacting this statute. The court therefore concludes that plaintiffs have failed to prove that the one-location rule violates the Fifteenth Amendment. That leaves the provisions that reduce the days and hours in which in-person absentee voting is allowed. Plaintiffs have adduced evidence that weekend and evening voting is particularly important for socioeconomically disadvantaged voters, and that, in Wisconsin and nationwide, African American and Latino voters have made particularly good use of various forms of early voting. See, e.g., PX036, at 42; PX047. Early voting in groups on Sundays—including church-supported “Souls to the Polls” efforts—is a widespread practice among African American voters, in Wisconsin and nationwide. Tr. 1p, at 134:6-135:1; PX245, at 38. But again, a disparate impact, without more, does not prove intentional discrimination. But plaintiffs have more. Statements by legislators show that Act 146 reduced the hours allowed for in-person absentee voting specifically to curtail voting in Milwaukee, and, secondarily, in Madison. Senator Grothman made repeated statements objecting to the 42 extended hours for in-person absentee voting in Milwaukee and Madison, indicating that hours for voting needed to be “reined in.”12 On the floor of the senate, he said, “I want to nip this in the bud before too many other cities get on board.” PX022, at 5. Senate Majority Leader Scott Fitzgerald made similar comments. Id. at 12. As he put it, “But the question of where this is coming from and why are we doing this and why are we trying to disenfranchise people, I mean, I say it’s because the people I represent in the 13th district continue to ask me, ‘What is going on in Milwaukee?’” Id. at 16. Defendants contend that Grothman and Fitzgerald were simply trying to achieve a measure of statewide uniformity because smaller towns were unable to afford the extended hours that Milwaukee was offering. That explanation is hard to credit. Under Act 146, the legislature still tolerates disparities in voting hours among Wisconsin municipalities. Each municipality can set its own hours for in-person absentee voting. Larger cities can still outdo smaller municipalities by having their full-time clerks hold office hours that cover the full work week, while smaller towns with part-time clerks will hold limited hours, sometimes as little as an afternoon a week. Thus, rather than achieving uniformity, the provisions governing the hours for in-person absentee voting preserved great disparities from town-totown. The legislative record shows that Act 146 was uniformly opposed by municipal clerks. PX216. Its only supporter of record was the Republican election activist Ardis Cerny. Id. And 12 Plaintiffs have adduced evidence that might suggest personal bias on Grothman’s part. PX078 (statements about Martin Luther King, Jr. Day); PX073 (about Milwaukee voters who would not be able to vote on weekends: “[A]nybody who can’t vote with all these options, they’ve really got a problem. I really don’t think they care that much about voting in the first place, right?”). The court does not ascribe Grothman’s personal antagonism toward minority voters to the legislature. 43 Governor Walker partially vetoed the bill as too extreme a reduction in opportunities to vote. PX058. The acknowledged impetus for this law was the sight of long lines of Milwaukee citizens voting after hours. Yet instead of finding a way to provide more access to voters in small towns, the legislature responded by reining in voters in Milwaukee, the state’s most populous city, where two-thirds of its African American citizens live. At trial, Kevin Kennedy, director of the GAB, confirmed that the purpose of reducing the hours for in-person absentee voting was to restrain voting in Milwaukee: Clearly in the recall election, the City of Milwaukee opened its in-person absentee voting for Memorial Day, which was the day before the gubernatorial recall election, and that did not sit well with the Republican majority. They thought that was designed purposely . . . to allow more Democratic voters, even though it could also be said it was designed to facilitate the needs of the unique voters in Milwaukee. But that was not lost on the Legislature that the largest city made that choice whereas other municipalities wouldn’t make that choice. Tr. 5a, at 109:21-110:5. The legislature’s ultimate objective was political: Republicans sought to maintain control of the state government. But the methods that the legislature chose to achieve that result involved suppressing the votes of Milwaukee’s residents, who are disproportionately African American and Latino. The legislature did not act out of pure racial animus; rather, suppressing the votes of reliably Democratic minority voters in Milwaukee was a means to achieve its political objective. But that, too, constitutes race discrimination. Ketchum v. Byrne, 740 F.2d 1398, 1408 (7th Cir. 1984) (“We think there is little point for present purposes in distinguishing discrimination based on an ultimate objective of keeping certain incumbent whites in office from discrimination borne of pure racial animus.”); see also Rogers, 458 U.S. at 44 617 (“[M]ultimember districts violate the Fourteenth Amendment if ‘conceived or operated as purposeful devices to further racial discrimination’ by minimizing, cancelling out or diluting the voting strength of racial elements in the voting population.”). Based on the evidence that plaintiffs have presented, the court finds that Wisconsin’s restrictions on the hours for in-person absentee voting have had a disparate effect on African Americans and Latinos. The court also finds that the legislature’s justification for these restrictions was meager, and that the intent was to secure partisan advantage. Finally, the court finds that the legislature specifically targeted large municipalities—Milwaukee in particular—intending to curtail minority voting. Combined, these findings lead the court to further find that the legislature passed the provisions restricting the hours for in-person absentee voting motivated in part by the intent to discriminate against voters on the basis of race. 2. Age discrimination Plaintiffs contend that some of the challenged provisions discriminate against younger voters on the basis of age, in violation of the Twenty-Sixth Amendment. The Twenty-Sixth Amendment provides that “[t]he right of citizens of the United States, who are eighteen years of age or older, to vote shall not be denied or abridged by the United States or by any State on account of age.” The federal courts that have considered Twenty-Sixth Amendment claims recognize that there is “a dearth of guidance on what test applies to Twenty-Sixth Amendment claims.” N.C. State Conference of the NAACP v. McCrory, No. 13-cv-658, 2016 WL 1650774, at *165 (M.D.N.C. Apr. 25, 2016), rev’d, No. 16-1468 (4th Cir. July 29, 2016)13; see also Walgren v. 13 The court has reviewed the Fourth Circuit’s decision invalidating North Carolina’s voter ID 45 Bd. of Selectmen of Amherst, 519 F.2d 1364, 1367 (1st Cir. 1975) (“[W]e are still without the assistance of any precedents guiding us in evaluating the impact of the Twenty-sixth Amendment.”); Nashville Student Org. Comm. v. Hargett, No. 15-cv-210, 2015 WL 9307284, at *6 (M.D. Tenn. Dec. 21, 2015) (“As the parties note in their briefing, there is no controlling caselaw from the Sixth Circuit or the Supreme Court regarding the proper interpretation of the Twenty-Sixth Amendment or the standard to be used in deciding claims for Twenty-Sixth Amendment violations based on an alleged abridgment or denial of the right to vote.”). The text of the Twenty-Sixth Amendment is patterned on the Fifteenth Amendment, which prohibits the denial or abridgement of the right to vote on the basis of race. This suggests that Arlington Heights provides the appropriate framework for evaluating plaintiffs’ claims of intentional age discrimination. Indeed, other courts have taken this approach when confronted with similar allegations. See, e.g., Lee v. Va. State Bd. of Elections, No. 15-cv-357, 2016 WL 2946181, at *26 (E.D. Va. May 19, 2016). Although the district court in North Carolina State Conference of the NAACP expressed doubt that the Twenty-Sixth Amendment was intended to operate just like the Fifteenth Amendment, the court followed an Arlington Heights-style analysis for the purposes of its decision. 2016 WL 1650774, at *165. Anderson-Burdick provides a framework through which the court could evaluate the burdens that fall on younger voters and the state’s justification for those burdens. But “[i]t is difficult to believe that [the Twenty-Sixth Amendment] contributes no added protection to that already offered by the Fourteenth Amendment, particularly if a significant burden were law on the grounds that it was motivated by an intent to discriminate on the basis of race. The decision relies on factual considerations unique to North Carolina, and, accordingly, it has no bearing on this case. 46 found to have been intentionally imposed solely or with marked disproportion on the exercise of the franchise by the benefactors of that amendment.” Walgren, 519 F.2d at 1367. Thus, for plaintiffs’ age discrimination claims, the court will apply the Arlington Heights framework, beginning by considering whether plaintiffs have shown that the challenged provisions have had a disparate impact on younger voters. All of the challenged provisions are facially neutral, but plaintiffs have offered anecdotal evidence that some of them disproportionately affect younger voters. See generally Dkt. 207, at 236-41 (discussing trial evidence). As a class, younger voters are poorer and less established. They are therefore less likely to have a driver license and documentary proof of residence. They are also more transient, and thus will likely face the burden of registration more often. But this evidence falls short of showing that young people are more likely to face burdens that they cannot overcome with reasonable effort. Young people may be more likely to lack a driver license. But that does not show that they are more likely to lack the credentials that one needs to get a Wisconsin ID. Young people may move more often, and they may be more likely to conduct their affairs online. But that does not mean that they will lack the documents needed to register, particularly because online documents can serve as proof of residence. The court does not find strong evidence of a disparate impact, which puts plaintiffs’ Twenty-Sixth Amendment claim on weak footing. Plaintiffs have some evidence of anti-youth comments made by legislators, particularly those by Senate Majority Leader Mary Lazich. Before the vote on Act 23, Lazich told the senate Republican caucus that they should support the bill because of what it “could mean for the neighborhoods of Milwaukee and the college campuses across this state.” Tr. 1a, at 84:1-24. As the court has already concluded, the Republican majority was motivated, in part, 47 by partisan objectives. But without more, this type of evidence did not establish discrimination on the basis of race, and it does not establish discrimination on the basis of age either. Much of plaintiffs’ evidence concerns the restrictions that the legislature placed on the use of college IDs. The rationale for these restrictions is not as weak as the rationale for the reduction in hours for in-person absentee voting. Under Anderson-Burdick, the court will evaluate whether these restrictions impose burdens that are warranted in light of the interests that they serve. But in the context of intentional age discrimination, the question is more limited: were these restrictions so baseless as to suggest purposeful discrimination against young voters? The court concludes that the answer is “no.” The restrictions served a legitimate interest in election integrity because many college students have documentation of two residences: their school addresses, and their permanent home addresses. The legislature had a legitimate interest in ensuring that students registered in only one place. See, e.g., PX229 (legislative note expressing interest in tightening up registration requirements so that out-of-state students would have to declare residency in Wisconsin to vote in the state). The court will review the state’s rationales for the other challenged restrictions later in this opinion. For the purposes of plaintiffs’ age discrimination claim, however, it is sufficient to say that these rationales are not so feeble as to suggest intentional discrimination. One last point. College students may use any of the means of identification or proof of residence that are available to all citizens generally. The legislature also extended to students the additional ability to use their college IDs, albeit under certain restrictive conditions. As a practical matter, these restrictions meant that the standard student IDs that many University of Wisconsin campuses issue were not valid for voting. But some 48 universities have provided workarounds in the form of special university-issued voting IDs. This seems like an unwarranted rigmarole, but the end result is that college students have more ID options than other citizens do. The court concludes that plaintiffs have not proven by a preponderance of the evidence that the challenged provisions were motivated by intentional age discrimination. D. Partisan fencing claim At the heart of this case is plaintiffs’ contention that the Wisconsin legislature passed the challenged provisions with the intent to suppress Democratic votes to gain a partisan advantage in future elections. Plaintiffs contend that to accomplish this objective, the legislature identified groups of voters who would likely vote for Democrats and then passed measures to frustrate those voters’ access to the ballot box. Put differently, the legislature targeted minorities, younger citizens, and citizens in urban areas like Milwaukee, not necessarily because of racial or age-based animus, but because it believed that these groups tended to vote for Democrats. Plaintiffs bundle these allegations into a “partisan fencing” claim. Dkt. 141, ¶¶ 197-99. This is not the first time that a group of plaintiffs in a voting rights case has asserted a partisan fencing claim. See, e.g., Lee v. Virginia State Bd. of Elections, No. 15-cv-357 (E.D. Va. filed June 11, 2015); Ohio Org. Collaborative v. Husted, No. 15-cv-1802 (S.D. Ohio filed May 8, 2015). But the legal theory is still a novel one, and neither party directs the court to precedent—binding or otherwise—that definitively establishes a framework for analyzing partisan fencing claims. Plaintiffs extrapolate that their partisan fencing claim is essentially a claim for intentional discrimination, relying on statements in various Supreme Court decisions. They therefore urge the court to consider their evidence of partisan motivation by 49 using the Arlington Heights framework, which would lead the court to invalidate any election qualification that was motivated, even in part, by partisan objectives. Defendants contend that a partisan fencing claim is really just a unique species of an undue burden claim, for which the Anderson-Burdick framework is appropriate. Plaintiffs derive the term “partisan fencing” from Carrington v. Rash, a case in which the Supreme Court invalidated a Texas constitutional provision that prevented members of the United States armed forces from voting if they moved to Texas during their service. 380 U.S. 89, 89 (1965). The Court held that “‘[f]encing out’ from the franchise a sector of the population because of the way they may vote is constitutionally impermissible.” Id. at 94. But the Court decided Carrington well before Anderson v. Celebrezze, 460 U.S. 780 (1983), and Burdick v. Takushi, 504 U.S. 428 (1992), the two namesake cases for the Anderson-Burdick framework that courts now apply to evaluate whether voting regulations burden First and Fourteenth Amendment rights. Moreover, Carrington dealt with an outright prohibition on voting—service members who moved to Texas during their military service could not vote while they were in the armed forces. Id. at 89. And cases applying Carrington tend to involve outright prohibitions on the right to vote. See, e.g., Evans v. Cornman, 398 U.S. 419, 419-20 (1970) (Maryland citizens who lived on a federal reservation prohibited from voting because they were not residents of Maryland); Cipriano v. City of Houma, 395 U.S. 701, 702 (1969) (per curiam) (“[O]nly ‘property taxpayers’ [had] the right to vote in elections called to approve the issuance of revenue bonds by a municipal utility.”). Here, none of the challenged provisions categorically bar any citizen of Wisconsin from voting. For these reasons, Carrington is not directly on point here. 50 Looking toward more recent cases, at least one Justice of the Supreme Court has suggested that there would be First Amendment implications for state restrictions on voting that place burdens on voters because of their political views. See Vieth v. Jubelirer, 541 U.S. 267, 315 (2004) (Kennedy, J., concurring) (“If a court were to find that a State did impose burdens and restrictions on groups or persons by reason of their views, there would likely be a First Amendment violation, unless the State shows some compelling interest.”). Several years later, a unanimous Court noted that this suggestion was “uncontradicted by the majority in any of our cases.” Shapiro v. McManus, 136 S. Ct. 450, 456 (2015). But these decisions involved gerrymandering, which is not at issue in this case. The import of these cases is that analyzing a partisan fencing claim involves a balancing analysis under the First Amendment. And that is exactly what the Anderson-Burdick framework provides. The framework requires the court to identify the nature and severity of the burden that a given voting regulation creates and then weigh that burden against the state’s justification for it. Common Cause Ind. v. Individual Members of the Ind. Election Comm’n, 800 F.3d 913, 917 (7th Cir. 2015). Thus, Anderson-Burdick appears to fit the bill for plaintiffs’ partisan fencing claim. Two federal district courts that have confronted this question reached the same conclusion. In Ohio Organizing Collaborative v. Husted, the Southern District of Ohio concluded that Carrington does not “appear to create a separate equal protection cause of action to challenge a facially neutral law that was allegedly passed with the purpose of fencing out voters of a particular political affiliation.” No. 15-cv-1802, 2016 WL 3248030, at *48 (S.D. Ohio May 24, 2016). Instead, the court relied on the Anderson-Burdick framework as “the proper standard under which to evaluate an equal protection challenge to laws that 51 allegedly burden the right to vote of certain groups of voters.” Id. Likewise, in Lee v. Virginia State Board of Elections, the Eastern District of Virginia acknowledged that “[t]he term ‘partisan fencing’ is derived from Carrington . . . and is somewhat of an aberration.” 2016 WL 2946181, at *26. The court concluded that the term “has been rarely deployed in election law litigation thereafter. It does not appear to create a separate cause of action but may be a useful analytical tool in evaluating First Amendment and Equal Protection Clause cases.” Id. The reasoning in these decisions is persuasive, and this court will follow their guidance. The court will not adopt plaintiffs’ partisan fencing theory, but the theory is not completely without basis. This case challenges state laws governing voter qualifications and election mechanics; it is not a redistricting case. That distinction is important. The redistricting process is inherently political through and through, and a gerrymandering claim requires a court to decide how much partisan politics is too much. See generally League of United Latin Am. Citizens v. Perry, 548 U.S. 399, 413-23 (2006). By contrast, voter qualifications and election administration should not be political at all, and partisan gain can never justify a legislative enactment that burdens the right to vote. So, plaintiffs argue, a state should not be allowed to manipulate its election regime by imposing even slight burdens, if the purpose is to suppress turnout to achieve a partisan advantage. Despite the appeal of plaintiffs’ theory, Crawford and Frank foreclose the argument that partisan fencing claims should be handled like claims of intentional race or age discrimination, for which any discriminatory legislative intent is sufficient to invalidate a law. See Frank, 768 F.3d at 755 (“‘[I]f a nondiscriminatory law is supported by valid neutral justifications, those justifications should not be disregarded simply because partisan interests may have provided one motivation for the votes of individual legislators.’” (quoting Crawford, 52 553 U.S. at 204)). Put differently, a provision is not unconstitutional if the legislators who passed it were partly motivated by partisan gain, so long as there were sufficient valid justifications. The Anderson-Burdick framework enables federal courts to undertake this type of review. In sum, the court rejects plaintiffs’ proposal to treat their partisan fencing claim as distinct from their undue burden claims under the First and Fourteenth Amendments. As explained below, the evidence of partisan motivation that plaintiffs have adduced is pertinent to the legislature’s justifications for passing the challenged provisions. The court will therefore consider this evidence as part of its Anderson-Burdick balancing analysis. E. First and Fourteenth Amendment claims for undue burdens on the right to vote Plaintiffs contend that each of the challenged provisions violates the First and Fourteenth Amendments by impermissibly burdening the right of Wisconsin citizens to vote. “A state election law, ‘whether it governs the registration and qualifications of voters, the selection and eligibility of candidates, or the voting process itself, inevitably affects—at least to some degree—the individual’s right to vote and his right to associate with others for political ends.’” Common Cause Ind., 800 F.3d at 917 (quoting Anderson, 460 U.S. at 788). But that is not to say that every voting-related law must survive strict scrutiny. Requiring states to narrowly tailor their election regulations to advance only compelling interests “would tie the hands of States seeking to assure that elections are operated equitably and efficiently.” Burdick, 504 U.S. at 433. Federal courts must therefore apply a “more flexible standard” when reviewing challenges to a state’s election laws. Common Cause Ind., 800 F.3d at 917. Under the flexible Anderson-Burdick standard, “the rigorousness of [the] inquiry into the propriety of a state election law depends upon the extent to which a challenged regulation 53 burdens First and Fourteenth Amendment rights.” Burdick, 504 U.S. at 434. The court must undertake a three-step analysis for each of the challenged provisions. First, the court must determine the nature and severity of the burden that a given provision imposes. Second, the court must identify the state’s justification for the provision. Third, the court must weigh the burdens against the state’s justifications for imposing them “and then make the ‘hard judgment’ that our adversary system demands.” Crawford, 553 U.S. at 190. For the first step in the Anderson-Burdick analysis, the court must focus on the burdens that the challenged provisions place on eligible voters who cannot comply with the new requirements (e.g., who lack registration documents, who need to vote during a different inperson absentee voting period or at a different location, or who prefer to vote straight-ticket). See id. at 198 (“The burdens that are relevant to the issue before us are those imposed on persons who are eligible to vote but do not possess a current photo identification that complies with the requirements of SEA 483.”). Just because the majority of Wisconsin voters are able to comply with the state’s registration requirements, absentee voting procedures, and miscellaneous election regulations does not mean that the burdens that these laws impose are constitutionally insignificant. But just as important, the fact that a few Wisconsin voters have difficulty complying with these laws is not enough to invalidate them across the board. Crawford, 553 U.S. at 199-200 (“And even assuming that the burden may not be justified as to a few voters, that conclusion is by no means sufficient to establish petitioners’ right to the [facial] relief they seek in this litigation.”). For the second step in the Anderson-Burdick analysis, the court must “consider the precise interests put forward by the State as justifications for the burden imposed by its rule, taking into consideration the extent to which those interests make it necessary to burden the 54 plaintiff’s rights.” Common Cause Ind., 800 F.3d at 921 (citations and internal quotation marks omitted). For the third step in the Anderson-Burdick analysis, the court must weigh the burdens of a given provision against the state’s justification for it. When the state imposes a “severe” restriction on the right to vote, then “the regulation must be narrowly drawn to advance a state interest of compelling importance.” Burdick, 504 U.S. at 434 (citations and internal quotation marks omitted). “But when a state election law provision imposes only ‘reasonable, nondiscriminatory restrictions’ upon the First and Fourteenth Amendment rights of voters, ‘the State’s important regulatory interests are generally sufficient to justify’ the restrictions.” Id. (quoting Anderson, 460 U.S. at 788). With these considerations in mind, the court turns to the specific provisions that plaintiffs challenge in this case. 1. Limiting in-person absentee voting In 2005, Wisconsin enacted Wis. Stat. § 6.855, which limited municipalities to one location for in-person absentee voting. At that time, the state did not limit the hours for inperson absentee voting. But as a practical matter, in-person absentee voting could not begin until municipal clerks received the ballots from the company that printed them, which was usually three to five weeks before the election. Tr. 2, at 265:5-7; Tr. 4p, at 121:3-11; Tr. 7a, at 114:9-15. Through Act 23, passed in 2011, and Act 146, passed in 2014, the legislature narrowed the window for in-person absentee voting to 10 days and prohibited municipal clerks from offering in-person absentee voting on weekends or on the Monday before an election. The legislature also limited the hours available for in-person absentee voting to between 8:00 a.m. and 7:00 p.m. 55 The court finds that the challenged in-person absentee voting provisions place a moderate burden on the right to vote. Wisconsin’s changes to its in-person absentee voting regime came amidst an increase in the use of absentee voting, both nationally and in Wisconsin. About 60,000 voters cast inperson absentee ballots on the Monday before the November 2008 general election. PX435, at 13. As plaintiffs’ expert, Barry Burden, PhD, testified, absentee voting in Wisconsin (both by mail and in-person) increased from 10.6 percent to 15.5 percent between the 2010 and 2014 midterm elections. PX037, at 23. For presidential elections, the increase was not as significant: 21.1 percent in 2008 to 21.4 percent in 2012. Id. Defendants’ expert, Dr. Hood, reached similar conclusions. Tr. 8a, at 32-41; DX001, at 11. In spite of these trends, plaintiffs contend that the one-location rule and hour limit stifled in-person absentee voting in Wisconsin. Their theory is that if the legislature had not passed the challenged provisions, then in-person absentee voting would have increased even more, particularly among minorities and young voters, who tend to vote for Democrats. The court agrees with Dr. Hood that it would be nearly impossible to directly prove this theory— there is no way to redo the 2012 and 2014 elections without the in-person absentee provisions in place. Tr. 8a, at 44:3-6. Neither side had compelling statistical evidence that African Americans in Wisconsin had made disproportionate use of in-person absentee voting. But plaintiffs had good anecdotal and circumstantial evidence that the in-person absentee laws impose burdens for certain voters by demonstrating that the changes had profound effects in larger municipalities like Madison and Milwaukee. These cities are home to populations of voters who disproportionately lack the resources, transportation, or flexible work schedules necessary to vote in-person absentee during the decreased timeframe. PX037, 56 at 26-27. At trial, clerks from both cities testified that the new laws forced them to drastically cut back on the amount of time that they could offer in-person absentee voting. For example, before the November 2012 elections, Madison offered in-person absentee voting until 8:00 p.m. on weekdays, and for a few hours on Saturdays and Sundays. Tr. 2, at 265:16-20. Up to 1,200 voters a day would use in-person absentee voting. Id. at 266:1-6. As for Milwaukee, defendants’ own expert summarized how the changes have similarly affected the availability of in-person absentee voting since 2008. DX001, at 9. Voters in both municipalities took advantage of the opportunities available before the state limited in-person absentee voting, particularly weekend voting. PX206. In Wisconsin, voters in larger cities experience disadvantages in education, income, employment, and access to transportation. PX036, at 5-15; PX037, at 26-27. Several lay witnesses testified that these pre-existing disadvantages interact with the new laws to make it more difficult for these voters to vote during the shorter period for in-person absentee voting. For example, eliminating weekend voting and reducing the number of days on which a clerk’s office can accept in-person absentee ballots is problematic for a person whose job or class 57 schedule is less flexible. Tr. 1p, at 14:13-15:8, 75:8-25, 144:19-25; Tr. 3p, at 31:2-5. Combined with the one-location rule, limiting hours leads to longer lines at clerk’s offices, which in turn requires voters to be prepared to devote more time to voting. Tr. 1p, at 92:1896:3; Tr. 2, at 266:7-16. Having only one location creates difficulties for voters who lack access to transportation. Eliminating weekend voting also prevented groups from holding voting drives like “Souls to the Polls”—an initiative that encouraged church congregations to vote in-person absentee after church on Sunday. Tr. 1p, at 134:20-135:1; Tr. 2, at 183:14-17. But these types of collateral effects only indirectly burden voters; impediments for groups trying to get individuals to vote do not necessarily implicate the First Amendment. Cf. Voting for Am., Inc. v. Steen, 732 F.3d 382, 388-96 (5th Cir. 2013) (“[W]e are unpersuaded that the smorgasbord of activities comprising voter registration drives involves expressive conduct or conduct so inextricably intertwined with speech as to require First Amendment scrutiny.”); Coal. for Sensible & Humane Sols. v. Wamser, 771 F.2d 395, 400 (8th Cir. 1985) (acknowledging the claim that “refusal to appoint qualified volunteers as deputy registrars restricts the accessibility of voter registration facilities and thus indirectly constitutes an unconstitutional infringement of the right to vote,” but refusing to “agree that there is a constitutional right to greater access to voter registration facilities per se”). The challenged provisions do not categorically bar individuals from voting. The state has shrunk the window in which municipalities can offer in-person absentee voting, but it has not closed that window completely. If the shortened period is not convenient for certain voters, then they can vote using mail-in absentee voting or vote on election day. Regardless, both sides’ evidence confirms that in-person absentee voting is still widely used, and its use 58 has increased over the last several years. As noted above, plaintiffs argue that without the challenged provisions, in-person absentee voting would be increasing more. But their anecdotal evidence is not sufficient to prove this assertion. Before turning to step two of the Anderson-Burdick analysis, the court will address defendants’ preliminary argument that there is no constitutionally protected right to cast an absentee ballot. Defendants invoke Griffin v. Roupas, a case in which a group of working mothers challenged Illinois’s refusal to let them vote absentee because they did not satisfy any of the statutory prerequisites (out of the county, physical incapacity, religious observance, etc.). 385 F.3d 1128, 1129 (7th Cir. 2004). The Griffin court rejected the idea “that the Constitution requires all states to allow unlimited absentee voting,” id. at 1130, which defendants implicitly contend should end the discussion. But this case is not about Wisconsin’s outright refusal to allow in-person absentee voting. Rather, plaintiffs allege that the state is denying them the opportunity to exercise a right that they already have. Put differently, plaintiffs contend that by choosing to give its citizens the privilege of in-person absentee voting, the state must administer that privilege evenhandedly. See Zessar v. Helander, No. 05-cv-1917, 2006 WL 642646, at *6 (N.D. Ill. Mar. 13, 2006) (“[O]nce [states] create such a regime, they must administer it in accordance with the Constitution.” (citing Paul v. Davis, 424 U.S. 693, 710-12 (1976))). The court therefore rejects defendants’ argument that plaintiffs’ challenge to the in-person absentee voting provisions does not implicate their constitutional rights. Defendants advance four justifications for the challenged in-person absentee voting provisions. First, they contend that shortening the timeframe for in-person absentee voting will allow the state to conduct uniform, orderly elections. Municipal clerks can better control 59 the process and manage staffing. Clerks can also guarantee that absentee ballots will be available once in-person absentee voting starts (ballots are delivered at different times, which means that a clerk’s office might have them available four weeks before an election one year, but only two weeks before that same election in a different year). Second, defendants contend that municipal clerks are busy during election season. With the reduced window for in-person absentee voting, clerks have more time for other tasks, such as conducting voting at residential care facilities, mailing absentee ballots, and entering voter registrations. Clerks also have non-election-related duties, and it becomes difficult to attend to them during business hours once in-person absentee voting begins. The reduced window allows them to take care of other responsibilities before turning their exclusive attention to voting. Third, defendants contend that limiting in-person absentee voting to one location saves money. More locations mean more staff, supplies, and security. Clerks are also able to directly supervise the entire process because it is occurring in one location rather than across the municipality. Fourth, defendants contend that limiting in-person absentee voting to one location avoids voter confusion by creating uniformity. Their concern is that voters might accidentally believe that because they can vote in-person absentee at multiple locations, they can also vote at multiple polling locations on election day. With one exception, these interests do not justify the moderate burdens that the challenged provisions impose. Alleviating the workload for clerks could be a sufficient reason to limit the hours for in-person absentee voting. But the laws that the challenged provisions replaced did not require municipal clerks to offer in-person absentee voting during the now- 60 eliminated days and times or at multiple locations. A clerk who wanted to retain control over the process, save money by using less staff, or reduce the hours to have time to attend to other duties could have chosen to do so under the old laws. Thus, any burdens on clerks that the state was purporting to address were voluntarily undertaken, which undermines the state’s interest in alleviating those burdens. Furthermore, the state’s interest in establishing uniform times for in-person absentee voting does not make sense because clerks can currently set whatever hours and days they want for in-person absentee voting, within the parameters of the statutes. Contrary to defendants’ assertion, Dkt. 206, at 65, the new laws do not actually “provide[] a set date when in-person absentee voting begins.” Municipal clerks are still free to start in-person absentee voting at different times, so long as it is not before the window opens. Under the new law, smaller towns with part-time clerks can still conduct in-person absentee voting by appointment only or on just a few days a week, see, e.g., Tr. 7a, at 166:21-177:14; PX161, while larger municipalities can offer in-person absentee voting from 8:00 a.m. to 7:00 p.m., Monday through Friday, for two weeks, see, e.g., Tr. 2, at 265:2-12. Thus, the challenged provisions do not actually create any consistency in when individual clerk’s offices offer inperson absentee voting. Requiring all municipalities to have one location for in-person absentee voting may have a superficial appeal. But uniformity for uniformity’s sake gets the state only so far. In 2014, the number of adults per municipality in Wisconsin ranged from 33 to 433,496. PX037, at 26. The state’s one-location rule ignores the obvious logistical difference between forcing a few dozen voters to use a single location and forcing a few hundred thousand voters to use a single location. There is simply no evidence that a one-location rule prevents voter 61 confusion, or that any confusion would be as widespread or burdensome as the types of difficulties that voters face when having only one location at which to vote in-person absentee. Evidence at trial suggested that one of the justifications for the challenged in-person absentee provisions was to “rein in” the big cities in the state, principally for political purposes. See generally PX022. State legislators were concerned that smaller municipalities could not keep up with the cities that had the resources to provide 60 to 70 hours of inperson absentee voting each week. Id. Ensuring equal access to the franchise is certainly a valid state interest, probably even a compelling one. But stifling votes for partisan gain is not a valid interest. And Wisconsin’s approach in this instance was backward: rather than expanding in-person absentee voting in smaller municipalities, the state limited in-person absentee voting in larger municipalities. By doing so, the state has imposed moderate burdens on the residents of those larger municipalities. The court concludes that most of the challenged in-person absentee voting provisions violate the First and Fourteenth Amendments for three reasons: the moderate burdens that they impose are not justified by the state’s proffered interests; local control addresses the needs of the communities; and the purported consistency is illusory. The one exception is the state’s decision to prohibit in-person absentee voting on the Monday before an election. The Wisconsin Municipal Clerks Association advocated for this provision, emphasizing that the day before an election is usually very busy. Tr. 4p, at 123:8124:12; Tr. 7a, at 158:22-160:9. The GAB advocated for this provision as well. Tr. 5a, at 102:2-4. The state’s interest in preventing clerks from incurring additional responsibilities on the day before an election, even voluntarily, is considerably more important than during the 62 weeks leading up to the election. Clerks cannot complete some of their preparation for election day until all absentee ballots are cast, and so allowing in-person absentee voting right up through the eve of the election necessarily prevents clerks from completing those tasks until after hours. Prohibiting in-person absentee voting on the day before an election allows clerks to focus on preparing for the election, go home at a reasonable hour, and be as sharp as possible for election day, which will itself be a long day. The state’s interest in prohibiting inperson absentee voting on the day before an election outweighs the moderate burdens that this measure imposes. Thus, the court concludes that this one provision does not violate the First and Fourteenth Amendments. 2. Requiring documentary proof of residence and eliminating corroboration Wisconsin requires voters to provide documentary proof of residence when registering to vote. Wis. Stat. § 6.34(2). Before Act 23, passed in 2011, voters could use corroboration to prove their residence. And before Act 182, passed in 2014, voters needed to provide documentary proof of residence only when registering to vote within 20 days before an election. Plaintiffs challenge both the requirement of documentary proof of residence and the elimination of corroboration. These are two aspects of an overall challenge to what Wisconsin requires from voters who want to register. Plaintiffs contend that Wisconsin’s proof of residence requirement burdens Wisconsin voters, particularly young voters who live with their parents, elderly voters, economically disadvantaged voters who live with friends or relatives, women voters whose residency documents are in their husbands’ names, and minority voters who suffer from higher rates of residential instability.14 14 Plaintiffs also contend that Wisconsin’s registration requirements have effectively put an end to voter registration drives. As explained above, the court’s primary task under AndersonBurdick is to evaluate the burden that a given provision places on voters. “[T]here is nothing 63 The court finds that the challenged registration provisions impose only slight burdens on voters. Between 2006 and 2012, about 35,000 Wisconsin citizens used corroboration to register to vote. PX038, at 39. But plaintiffs have adduced only anecdotal evidence to support their contention that the elimination of corroboration imposes a severe burden. They have not proven that minorities, Democrats, or young voters experience any widespread or insurmountable difficulties registering to vote on account of this change in the law. Indeed, plaintiffs’ expert conceded that he did “not have specific data on how many people were unable to register because they were no longer permitted to use corroborating witnesses to prove residency.” Id. The same is true of plaintiffs’ evidence about voters who could not provide documentary proof of residence: although plaintiffs have identified examples of voters who were turned away at the polls, there is no evidence about how prevalent the problem is, or about how many voters cannot obtain documentary proof of residence with reasonable effort. Voters in Wisconsin can satisfy the proof of residence requirement with a little planning. For example, rather than trying to register on election day, voters can contact their municipal clerk beforehand, when there is still time to update mailing addresses for bank statements, utility bills, or other acceptable forms of proof of residence. See PX490, at 5-6 (voter tried to use corroboration at the polls); PX045, at 3 (same); PX059, at 1 (183 people not able to register at polls because they did not have proof of residence). Wisconsin also ‘inherently expressive’ about receiving a person’s completed application and being charged with getting that application to the proper place,” Voting for Am., Inc., 732 F.3d at 392 (citations omitted), which means that the First Amendment would not protect a group’s mere desire to register voters. Plaintiffs’ evidence regarding voter registration drives is mostly tangential to the main issues in this case. 64 allows voters to present electronic copies of their proof of residence documents (e.g., online bank statements or utility bills), which eliminates the need to wait for a document to arrive by mail. At least some clerks have even identified a solution for voters who are simply unable to obtain the necessary documentation. Under Wis. Stat. § 6.34(3)(a)11., a person can register to vote by providing a document issued by a unit of government. Thus, if a voter provides a municipal clerk with the address at which the voter wants to register, the clerk can send the voter a letter and that letter then becomes a government document that the voter can use to register. See, e.g., Tr. 1p, at 163-65; Tr. 2, at 301-02. This system is not much different from the one that Wisconsin used to have. When a voter registered, the clerk’s office would send him or her a postcard to confirm the registration address. If the card came back as undeliverable, then the clerk’s office knew that there was a problem; if the card did not come back, then the clerk’s office considered the registration verified. The current laws merely add the step that a voter must return to the clerk’s office to verify receiving the document. The lone context in which proof of residence requirements and the elimination of corroboration can be more problematic is election day registration. An unregistered voter who lacks easy access to documentary proof of residence and decides on election day that he or she will vote may be unable to register without corroboration. The specific burdens on voters who plan to register on election day are still slight. With a little advanced planning, even a voter who lacks access to standard methods for proving residence can register to vote on election day. For many voters, registering to vote will not be a regular event: once registered, a voter can continue voting under that registration until he or she moves. And even for voters who 65 move often, if they complete the registration process once, they will be prepared for it in the future. Wisconsin law allows voters to choose from an array of documents to prove residence, and this flexibility means that the loss of corroboration does not impose a severe burden on the right to vote. It may be inconvenient to plan ahead to register at the polls on election day, particularly without corroboration, and it may be cumbersome to update account information with a bank or utility company. But these activities are no more burdensome than those that the Supreme Court has already considered. See Crawford, 553 U.S. at 198 (“For most voters who need them, the inconvenience of making a trip to the BMV, gathering the required documents, and posing for a photograph surely does not qualify as a substantial burden on the right to vote, or even represent a significant increase over the usual burdens of voting.”). Defendants justify the registration requirements as ensuring that voters actually reside in the municipalities where they register to vote. Asking for proof of residence, and not accepting corroboration, also helps prevent fraud. Defendants adduced no actual evidence of fraudulent use of corroboration though. See, e.g., Tr. 7a, at 118:20-119:6 (voter attempted to pressure other voters to corroborate his residence but they all refused). These interests justify the slight burdens that the challenged registration provisions impose. Residence is a bona fide voter qualification. Plaintiffs are correct that defendants have not adduced evidence of a genuine threat or history of registration-related fraud. But “[l]egislatures . . . should be permitted to respond to potential deficiencies in the electoral process with foresight rather than reactively, provided that the response is reasonable and does not significantly impinge on constitutionally protected rights.” Munro v. Socialist Workers Party, 479 U.S. 189, 195-96 (1986). Pursuant to Frank and Crawford, states can anticipate 66 and guard against fraudulent voting, and public confidence in elections is a legitimate state interest.15 Regardless, a voter’s residence in a particular municipality is a qualification for voting in that municipality. The state has an interest in making sure that only qualified voters are participating in elections, and the proof of residence requirement is directly linked to that goal. The court concludes that the challenged registration requirements do not violate the First and Fourteenth Amendments. 3. Changing how students can use “dorm lists” to register Before Act 23, college and university students could register to vote use their student IDs and a “dorm list” that their institutions provided to municipal clerks. 16 The legislature has changed this provision by requiring that dorm lists also indicate whether students are U.S. citizens. This change requires colleges and universities to provide information that the Family Educational Rights and Privacy Act (FERPA), 20 U.S.C. § 1232g, prevents them from disclosing without consent. PX435, at 34-35.17 Rather than obtaining consent to provide this information, most colleges and universities have stopped providing dorm lists to municipal clerks. PX436, at 10. 15 Frank and Crawford dealt with the requirement of presenting ID at the polls on election day. Presenting documentary proof of residence is the functional equivalent of a photo ID for the registration side of elections. 16 A dorm list is “a certified and current list of students who reside in housing sponsored by the university, college, or technical college.” Wis. Stat. § 6.34(3)(a)7.b. 17 FERPA permits colleges and universities to release only “directory information” without parental consent. This information includes “the student’s name, address, telephone listing, date and place of birth, major field of study, participation in officially recognized activities and sports, weight and height of members of athletic teams, dates of attendance, degrees and awards received, and the most recent previous educational agency or institution attended by the student.” 20 U.S.C. § 1232g(a)(5)(A). 67 The court finds that the dorm list provision places only a slight burden on student voters. The dorm list provision is a special accommodation that allows college and university students to prove their residences with student IDs. This option is in addition to the standard options that all voters have. Act 23 pulls back only some of the special dispensation that the legislature gave students. The challenged provisions do not deny students the ability to register outright. Students can also register using a student ID and a fee receipt showing that they paid tuition in the last nine months. See Wis. Stat. § 6.34(3)(a)7.a. And of course, students can register by presenting any of the other listed documents to prove residence. Plaintiffs did not present evidence showing how often students used dorm lists before Act 23, or how many students are now unable to register without the option. Without this sort of proof, plaintiffs cannot demonstrate that any burden on student voters is more than slight. Act 23 nevertheless burdens student voters who want to use their student IDs as proof of residence to register because it conditions their registration on proof of citizenship, which is something that no other voter must present to register. When any voter registers in Wisconsin, including a student voter, the voter must sign a statement certifying that he or she is a U.S. citizen. See DX101. But that is it. Voters do not need to actually prove that they are citizens. True, the primary burden that this provision imposes is on colleges and universities, which must provide compliant dorm lists. But if colleges and universities are unwilling to provide these lists, then for all practical purposes, Act 23 has taken away a method through which students can register to vote. Defendants justify the provision by arguing that U.S. citizenship is a qualification for voting in Wisconsin, see Wis. Const. art. III, § 1, and so “it makes sense to confirm it.” 68 Dkt. 206, at 87.18 That is a weak justification for two reasons. First, none of the state’s other methods for proving residence require voters to “confirm” their U.S. citizenship beyond signing a citizenship certification on the registration form. Students sign this certification too. Defendants do not explain how this certification procedure, which apparently satisfies the state’s interest in confirming citizenship for the overwhelming majority of non-students who register to vote, is insufficient in the context of student voters. Second, even if the state is particularly worried about non-citizen students voting—and at trial, the state presented no evidence of such a problem—the challenged provision does not allay that concern. Noncitizen students could easily skirt the requirement of demonstrating citizenship by using one of the other methods for proving residence. Although the changes to using a dorm list to register impose only slight burdens, the state has not offered even a minimally rational justification for the law. The court therefore concludes that this provision violates the First and Fourteenth Amendments. 4. Eliminating statewide SRDs and eliminating SRDs and registration locations at high schools Plaintiffs challenge the provisions of Act 23, passed in 2011, that eliminated statewide SRDs and the provisions of Act 240, passed in 2012, that eliminated the requirement that high schools accept registrations from staff and enrolled students. The court finds that the challenged SRD and high school registration provisions place only slight burdens on voters. 18 Defendants also argue that students have other options for proving residence. But that is not a justification for the law; as explained above, it is a reason for concluding that the law imposes only slight burdens on student voters. 69 Most of the burdens that plaintiffs identify from these laws do not fall directly on voters. For example, plaintiffs contend that eliminating statewide SRDs hinders individuals who register voters during off-site registration drives. See, e.g., Tr. 1p, at 7:20-8:25 (Citizen Action employee cannot register voters outside the municipalities in which she is an SRD), 187:15-188:6 (college student cannot be a statewide SRD); Tr. 3a, at 101:1-102:21 (organizations cannot conduct voter-registration drives). Plaintiffs also contend that without statewide SRDs, more voters will be forced to register at a municipal clerk’s office or at the polls, which will cause congestion and additional work for clerks and poll workers. Tr. 2, at 327:14-20. The Anderson-Burdick framework does not focus on these burdens; rather, the relevant issue is the nature and severity of the burdens that fall on voters and on the right to vote. The real burden for voters is the loss of potentially convenient options for registering through a statewide SRD or at a high school. But plaintiffs have not adduced evidence of how widespread or significant this problem is. No testimony or expert opinion established how many voters want to register through statewide SRDs or at high schools and are unable to do so. Nor did any testimony establish how many voters are unable to register at all without these options. The closest that plaintiffs came was an anecdote about one municipality not appointing any SRDs in 2011 and 2012, which meant that all voters had to register through the clerk’s office those years. PX490, at 3. Yet that burden was principally the result of that particular clerk refusing to appoint any SRDs. Plaintiffs do not argue that all, or even many, other municipalities refuse to appoint SRDs. Defendants justify these provisions by arguing that statewide SRDs make mistakes that municipal clerks have to spend time correcting. Tr. 4p, at 133:3-20 (continuous 70 difficulties in municipalities across the state with untimely or incorrect registrations from SRDs); Tr. 7a, at 121:2-7 (statewide SRDs submit incomplete forms, “which complicates things and requires follow-up”), 170:6-19 (same); Tr. 8p, at 133:8-12 (GAB auditor had problems with legibility and missing information from statewide SRDs). Defendants also presented evidence that students and staff did not use high school registration locations that frequently, and that high school SRDs also had problems submitting registrations. Tr. 4p, at 130:18-23 (problems with high school SRDs), 131:8-17 (less than 10 registrations per year from a high school), 132:3-9 (high school students like to register on election day or in the clerk’s office because “it’s a Facebook picture-taking time”); Tr. 7a, at 169: 11-19 (clerk has never received a registration from a high school and has not heard complaints about eliminating high schools as registration locations). Although this evidence was not conclusive for every municipality in the state, it supported defendants’ assertion that voters did not use high school registration locations that much. Plaintiffs counter these concerns by pointing out that they came only from small municipalities. Clerks from larger municipalities supported having statewide SRDs. Tr. 1p, at 88:3-8. Plaintiffs also argue that even if statewide SRDs make mistakes, these lead municipal clerks to engage with voters to correct those mistakes, and so the net result is beneficial. Plaintiffs’ criticisms are not persuasive: a state certainly does not have to stand by and watch problems fester in smaller municipalities just because one or two larger municipalities do not have, or can easily overcome, those same problems. The legislature was entitled to conclude that the problems with statewide SRDs outweighed the benefits. 71 Defendants also justify eliminating statewide SRDs on the grounds that it gave clerks direct control over the SRDs in their municipalities.19 The state supervised statewide SRDs, which made it difficult for municipal clerks to revoke or train SRDs when problems occurred. Tr. 4p, at 132:10-24. The benefits of local control led the Wisconsin Municipal Clerks Association to support eliminating statewide SRDs. Id. Now, clerks train and supervise each SRD in their municipality, which allows them to address issues quicker and more efficiently. The state’s interests in eliminating mistakes from high school and statewide SRDs, and in giving municipal clerks the ability to directly manage the SRDs with whom they work, justify the slight burdens that the challenged provisions impose. There is nothing stopping an individual from registering to be an SRD in as many municipalities as he or she likes. And alternative registration options alleviate virtually any inconvenience to voters who would benefit from being able to register with a statewide SRD. The court concludes that the challenged SRD and high school registration provisions do not violate the First and Fourteenth Amendments. 5. Preempting Madison’s landlord ordinance Act 76, passed in 2013, overrode an ordinance that Madison passed in July 2012 requiring landlords to distribute voter registration forms to new tenants. Plaintiffs contend that the act burdens the right to vote by making it harder to register. The court finds that the landlord provision imposes only a slight burden on voters. 19 The court notes that for this issue, the parties have switched sides on the importance of local control. Plaintiffs—for whom local control was so important in the context of in-person absentee voting—now appear to want statewide control, and defendants—for whom uniformity was so important in the context of in-person absentee voting—now argue that local control is vital. 72 There is some evidence that Madison’s ordinance was an effective tool for reaching voters who rented their homes. See, e.g., Tr. 3a, at 24:17-25:4. In the short time that Madison’s ordinance was in effect, Madison registered at least 500 voters who submitted the forms that their landlords had given them. Id. at 168:4-9. That was right before the November 2012 presidential election.20 Madison is also home to a large student population, with many students renting their homes. As with other challenged provisions, plaintiffs have not adduced evidence of a significant or widespread burden. The state statute does not preclude landlords from distributing materials; it just prevents municipalities from requiring that they distribute materials. Even assuming that in practice the law means that no landlord will provide forms, the only real burden that voters experience is having to obtain registration forms elsewhere— the rest of the steps for registering are the same. At most, the state has denied Madison voters a convenience. Plaintiffs have not adduced evidence of voters in Madison (or anywhere in Wisconsin) who did not receive registration forms from their landlords and were unable to register to vote. Defendants justify the law on the grounds that requiring landlords to provide voting materials creates the possibility for voter confusion. At trial, two municipal clerks opined that landlords, who are not trained election officials, could distribute outdated materials or inaccurate information. Tr. 4p, at 136:22-137:20; Tr. 7p, at 19:10-20:7. This testimony was speculative; defendants did not introduce evidence that landlords have actually distributed 20 The municipal clerk could not remember if it was the 2010 or 2012 election. But the ordinance went into effect in July 2012. See Madison, Wis., Code of Ordinances, § 32.06(5). 73 the wrong information. But the potential for confusion is at least plausible, which makes the state’s interest in avoiding it a reasonable one. The state has an interest in ensuring that voters receive the correct information about where and how to register to vote. Here, the possibility that landlords will provide outdated or inaccurate information seems minimal, and defendants’ justification for overriding Madison’s ordinance is relatively weak. If the statute more than minimally burdened the right to vote, then it probably would not withstand constitutional scrutiny. But defendants have put forth a rational explanation for it, and that explanation is sufficient to justify the slight burden that the law imposes. The court concludes that the landlord provision does not violate the First and Fourteenth Amendments. 6. Increasing the durational residency requirement Act 23, passed in 2011, increased Wisconsin’s durational residency requirement from 10 days to 28 days. This means that residents who move within Wisconsin fewer than 28 days before an election have to vote in their former municipalities. And residents who move into Wisconsin from out-of-state fewer than 28 days before an election cannot vote in Wisconsin at all (except for the offices of president and vice president, pursuant to Wis. Stat. § 6.15(1)). The court finds that the increased durational residency requirement imposes a moderate burden on voters in Wisconsin, particularly for populations that tend to be more transient or lack access to transportation. “Durational residence requirements completely bar from voting all residents not meeting the fixed durational standards. By denying some citizens the right to vote, such laws 74 deprive them of a fundamental political right, preservative of all rights.” Dunn v. Blumstein, 405 U.S. 330, 336 (1972) (citations, internal quotations, and alterations omitted). Plaintiffs have adduced evidence from which the court can infer that a longer residency requirement leads to increased difficulties for certain types of voters. That is an important consideration because the court must evaluate the burdens that the law imposes on voters who cannot comply with it. See Crawford, 553 U.S. at 198. Here, the burden is significant. A voter who does not satisfy the durational residency requirement cannot vote unless he or she: (1) travels back to his or her former municipality; or (2) votes absentee by mail. These options reduce the burden that the law imposes, but they do not negate it entirely. Plaintiffs seek a return to the old 10-day rule, presumably because the rule does not impermissibly burden the right to vote. Thus, their contention is really that the increase from 10 days to 28 days burdens the right to vote. Given the specific burdens at issue, plaintiffs’ evidence of problems with the overall durational residency requirement, see e.g., Tr. 1p, at 44:19-45:6; PX055, at 2; PX059, at 1, is not particularly relevant. Plaintiffs have not adduced direct evidence of the burdens that the change from 10 days to 28 days imposes. They have not identified how many voters would be able to comply with a 10-day rule but not with a 28-day rule. See Tr. 1p, at 44:9-14 (Citizen Action employee unable to identify how many voters were affected by the increase); Tr. 2, at 292:17-25 (municipal clerk testified to an unspecified “increase”); PX490, at 18 (one voter affected by the increase).21 Nor could plaintiffs’ experts pin down how widespread the problem is. For example, Dr. Lichtman presented 2010 census data to show that only 1.6 21 Defendants offered anecdotal evidence that not very many voters fall into the window between 10 and 28 days. See, e.g., Tr. 7a, at 122:4-10, 172:22-173:6. But this evidence, too, is inconclusive. 75 percent of the white population had moved into the state during the previous year, compared 2.1 percent of African Americans and 2.4 percent of Latinos. PX036, at 47. For in-state moves, 12.5 percent of white residents had lived in a different house in the previous year, compared to 26.2 percent of African Americans and 19.5 percent of Latinos. Id. at 41. But this information covered the entire year and was not limited to eligible voters. As with many of their other claims, plaintiffs attempted to indirectly prove the nature and severity of the burdens that the increased durational residency requirement creates. Voters who move more often have to confront residency requirements more often. Wisconsin has a significant population of African American and Latino voters, who are more likely to be transient than white voters are. PX036, at 40-41; PX037, at 27. Thus, the court can infer that the durational residency requirement will impose considerable burdens on a class of voters within the state that will have difficulty complying with the requirement. For voters who move into Wisconsin from another state, the 28-day residency requirement disenfranchises them from state and local elections in Wisconsin (although they can vote for president and vice president). Voters who move within the state at least have the option of voting in their former municipalities. But that option is realistically available only to those who can travel. Although voting absentee by mail can alleviate some of the burden for voters who cannot travel, that option presents its own obstacles. There is considerable public distrust of voting absentee by mail, the process is cumbersome and difficult to understand for some voters, and it presents added security challenges for municipal clerks. Tr. 1p, at 76:13-77:24; Tr. 2, at 114:18-117:10; Tr. 4p, at 158:7-159:14. On top of the burdens of actually voting in a former municipality, the durational residency requirement presents unique registration problems as well. Voters who must 76 register in their former municipalities may no longer have documents to prove their residence. Tr. 1p, at 79:16-22; Tr. 2, at 290:3-291:2. And even if a voter has adequate documentation, the registration form requires signing a certification that the voter has “resided at the [former] residential address for at least 28 consecutive days immediately preceding this election, with no present intent to move.” DX101, at 1. Signing this certification puts voters in an uncomfortable position because the form states that “[f]alsification of information on this form is punishable under Wisconsin law as a Class I felony.” Id.; see also Tr. 1p, at 79:7-15; Tr. 2, at 290:3-291:2. Also, for voters who sign the form and are able to register, there may still be confusion when the municipal clerk sends a confirmation postcard to confirm the new registration at the old address and the card is returned as undeliverable. PX436, at 24. Defendants justify the longer residency requirement as preserving election integrity, safeguarding voter confidence, and avoiding voter confusion. Specifically, the requirement serves these interests by preventing voter “colonization,” which “involve[s] voting by nonresidents, either singly or in groups. The main concern is that nonresidents will temporarily invade the State or county, falsely swear that they are residents to become eligible to vote, and, by voting, allow a candidate to win by fraud.” Dunn, 405 U.S. at 345. Defendants also contend that the requirement prevents “party raiding,” “whereby voters in sympathy with one party designate themselves as voters of another party so as to influence or determine the results of the other party’s primary.” Rosario v. Rockefeller, 410 U.S. 752, 760 (1973). Defendants’ purported interests in the 28-day durational residency requirement do not justify the severe burdens that the provision imposes for several reasons. First, defendants 77 did not introduce any evidence at trial of a genuine threat of colonization or party raiding. Nor have defendants explained how a durational residency requirement prevents party raiding, which is a problem that involves voters who are already registered. Second, even if the threat of colonization motivated the state’s actions, defendants failed to address the difference between a durational residency requirement in the abstract, and increasing that requirement from 10 days to 28 days. The state’s interests certainly justify some sort of residency requirement. See Marston v. Lewis, 410 U.S. 679, 680 (1973) (per curiam) (upholding a 50-day rule and holding that “[s]tates have valid and sufficient interests in providing for some period of time—prior to an election—in order to prepare adequate voter records and protect its electoral processes from possible frauds”). But defendants have not explained how a 28-day rule serves these interests better than a 10-day rule does. The court is not persuaded that increasing a durational residency requirement by 18 days actually inhibits colonization, raiding, or fraud, at least not to the extent necessary to justify the burdens that the increase imposes on otherwise-qualified voters. To the contrary, the requirement appears to simply make it harder for otherwise eligible voters to vote. It is also somewhat inconsistent with allowing election day registration, which lets voters decide to vote at the last minute. The state also advances a few practical points, which go toward avoiding voter confusion. For example, a GAB official testified that “the justification put forward to support the 28-day residency is partly that it was maybe more consistent with what some other states had.” Tr. 8p, at 41:16-18. Indeed, 25 states and the District of Columbia have a durational residency requirement, and the average length is 28.8 days. DX001, at 23. In 77 percent of those states, the requirement is 30 days. Id. The shortest requirement is 20 days. Id. at 24. 78 Consistency with other states is a superficial rationale that does not justify burdening (or completely disenfranchising) voters within the state who cannot comply with the requirement. Nor did defendants present evidence that there were such persistent problems with registration fraud (or any problems, for that matter) that the state needed to lengthen its durational residency requirement. Defendants also argue that the increased requirement allows voters more time to gather documents and plan for voting. For example, a voter who moves to a new district 11 days before an election might not have enough time to obtain documentary proof of the new residence, and a voter who moves 9 days before an election might not have enough time to request an absentee ballot from his or her former municipality. Any such convenience is utterly speculative—defendants did not identify a single voter who benefitted from the increased time in which to gather registration documents. Regardless, the rule adds considerable inconvenience. As one municipal clerk testified during trial, the rule is cumbersome for a person who moves 20 days before an election and is able to gather the necessary registration documents. Tr. 7a, at 140:16-142:1. Thus, defendants’ conveniencebased justification is not persuasive. The court concludes that the state’s change to the durational residency requirement violates the First and Fourteenth Amendments. 7. Establishing a zone for election observers Act 177, passed in 2014, established a statutorily prescribed zone in which election observers must stand at the polls to oversee voting on election day. The zone had to be between three and eight feet away from the table at which voters announced their names or registered to vote. Wis. Stat. § 7.41(2). This act overrode an existing GAB rule that allowed 79 observers to be between 6 and 12 feet from the location where voters were announcing their presence and registering to vote. Part of the impetus for Act 177 was that a select group of election observers complained that officials were invoking the GAB’s rule to keep them too far away to be able to hear and see events at polling places. See PX240; PX441, at 14-15. Plaintiffs allege that the state burdened the right to vote by moving observers closer to voters and facilitating harassment and intimidation. The court finds that the provisions governing where election officials can position election observers imposes only a slight burden on the right to vote. Although the executive director for Milwaukee’s Election Commission confirmed that “99.5% of election observers respect the state’s election observer rules,” Tr. 1p, at 112:16-18, some municipalities have had problems with disruptive, harassing, and intimidating observers. These problems are prevalent in high-minority areas like Milwaukee and Racine. PX045, at 3; PX436, at 19. Besides intimidating voters, having observers close to poll workers implicates voter privacy concerns: depending on the types of documents that a voter presents for registering or as identification, an observer could be able to see financial statements, social security numbers, or other personal information. Overly zealous election observers also potentially slow down poll workers and cause delays at the polls. Plaintiffs contend that these problems would not exist, or would at least not rise to the level of constitutional violations, under the GAB’s former 6-to-12-foot rule. Despite the evidence of problems with some observers, plaintiffs have not shown that Act 177 imposes a significant burden on voters. The court does not doubt that election observers can create consternation for many voters. But Wis. Stat. § 7.41(2) gives municipal clerks and chief election inspectors discretion to create an observation area at each polling 80 place; it does not require that they place observers closer than the GAB rule allowed. The court is not persuaded that the statute imposes any significant burden on voters. Local election officials have the discretion, under the statute, to manage the position of observers. In the anecdotes that plaintiffs presented at trial, problems with election observers occurred when poll workers or chief inspectors failed to exercise the authority that the state gave them to control or even remove observers. Problems also occurred when observers were closer than three feet, which was not a situation that the state even allowed, let alone imposed on voters. See, e.g., Tr. 1p, at 85:4-6 (“Well, to be clear, that wasn’t related to the space, the space issue; that was just related to the conduct of the observer.”). Also, plaintiffs’ evidence of problems consisted of incidents that occurred before the state passed Act 177, which undermines their assertion that the new law burdens the right to vote. Plaintiffs’ challenge to Wisconsin’s election observer law is essentially dissatisfaction with the choices that clerks or chief inspectors have made, or with their failure to address unruly observers. By establishing a range in which officials can place observers, the state has arguably made it possible for others to impose burdens on voters. But plaintiffs have failed to prove that election officials consistently exercise their authority under Wis. Stat. § 7.41(2) in a way that impedes or intimidates voters. At most, then, the law imposes only a slight burden on the right to vote. Defendants offer a compelling justification for giving municipal clerks and chief election inspectors discretion to establish an observation zone. “States may, and inevitably must, enact reasonable regulations of parties, elections, and ballots to reduce election- and campaign-related disorder.” Timmons v. Twin Cities Area New Party, 520 U.S. 351, 358 (1997). Here, the state balanced the right that observers have to be present at the polls with the 81 rights that voters have to keep their personal information private and with the flexibility that poll workers need to conduct efficient and fair elections. Rather than setting a one-size-fits-all rule, the legislature created guidelines to allow local municipalities to organize and control their polling places. Flexibility is important because not all polling places can accommodate a uniform distance. Tr. 2, at 286:17-289:22; Tr. 4p, at 139:18-140:2. And the range that the legislature selected was not unreasonable: three feet may be necessary to accommodate elderly observers or cramped polling places; eight feet allows observers to see and hear without interfering with poll workers. To be clear, the court does not condone harassment or intimidation by election observers, at any distance from registration or announcement tables. The state would be well served to impress upon municipal clerks and chief inspectors the importance of managing election observers. And those election officials must in turn exercise their authority to protect voters from unruly observers. As far as Act 177 is concerned, however, the state’s justification for the act outweighs any burdens that it creates. The court concludes that the challenged election observer provisions do not violate the First and Fourteenth Amendments. 8. Eliminating straight-ticket voting Act 23, passed in 2011, eliminated straight-ticket voting: voters must now select individual candidates on their ballots. Plaintiffs contend that this burdens the right to vote, particularly for voters with lower levels of educational attainment. The court finds that this provision creates only a slight burden on the right to vote, even among populations with lower levels of educational attainment or who have less time to spend voting. 82 The burdens that plaintiffs identify include longer lines at the polls (because voters must mark an entire ballot) and increased confusion and likelihood of mistakes. But there was limited evidence about whether the elimination of straight-ticket voting caused these burdens and, if so, to what extent. Dr. Lichtman wrote in his report that “[t]he elimination of straight-ticket voting in Act 23 also has an adverse impact on waiting time since it makes voting lengthier for those who would otherwise use this option.” PX036, at 44. Yet Dr. Lichtman did not identify evidence to support this assertion or indicate how much delay the elimination of straight-ticket voting actually caused. As for lay witnesses, plaintiffs elicited testimony that the lack of straight-ticket voting could confuse voters. See, e.g., Tr. 1p, at 82:17-83:3. But the actual evidence of confusion involved voters who remembered having the option in the past and asking about whether it still existed. PX490, at 22-23. Beyond that, straight-ticket voting was mostly a convenience, and plaintiffs did not adduce evidence that the lack of straight-ticket voting deterred anyone from voting. Defendants’ first justification for eliminating straight-ticket voting is that it was joining a national trend. As another district court recently explained, that argument does not get the state very far. Mich. State A. Philip Randolph Inst. v. Johnson, No. 16-cv-11844, 2016 WL 3922355, at *8 (E.D. Mich. July 21, 2016) (“The fact that some other states do not allow straight party voting changes none of the facts that are before this Court. Furthermore, and more importantly, the behaviors of other states are irrelevant to the question of constitutionality. If the Ohio Legislature successfully instituted poll-taxes and literacy tests without challenge, it would not change the fact that poll-taxes and literacy tests are still clearly unconstitutional burdens on the right to vote.” (original emphasis)). 83 Defendants also argue that eliminating straight-ticket voting decreases the chance of a voter selecting a straight-ticket option and then voting for candidates on the rest of the ballot. This type of over-voting would invalidate some or all of a voter’s choices. Wis. Stat. § 7.50(1)(b). Defendants did not introduce evidence that these types of problems were prevalent, although they seem no more or less likely than the confusion that some voters might experience after not seeing a straight-ticket option that they are used to. Nevertheless, defendants’ justification is reasonable. Finally, defendants argue that eliminating straight-ticket voting encourages voters to become more informed about candidates or issues, and it ensures that voters do not accidentally overlook items on a ballot. Defendants did not introduce evidence of how often these problems occur, but the danger is there: in elections with referenda or non-partisan races, a voter who uses a straight-ticket option could overlook some items on a ballot. Tr. 7p, at 20:8-21:23. This justification is reasonable. The court concludes that the straight-ticket provision does not violate the First and Fourteenth Amendments. 9. Prohibiting clerks from sending absentee ballots by fax or email Act 75, passed in 2011, prevents municipal clerks from faxing or emailing absentee ballots, except to military or overseas electors. Plaintiffs contend that this provision unjustifiably burdens voters who are traveling but who do not qualify as overseas electors. The court finds that this provision places a moderate burden on voters who are traveling, particularly if they are outside of the country or in locations with unreliable mail delivery. 84 Before Act 75, some municipalities sent hundreds of ballots by fax or email. Tr. 1p, at 87:8-12; Tr. 2, at 332:11-22. Now, without the option for electronic ballots, absentee voters must rely on mail service. This is particularly problematic for students or researchers who are abroad in remote areas, but it also affects domestic travelers, especially for elections in which ballots are not finalized until close to election day. Tr. 2, at 329:8-332:10; Tr. 7a, at 144:25145:23; PX491, at 6-9. In at least some cases, voters who cannot receive ballots by fax or email are simply unable to vote. Although voters are able to request their ballots by fax or email, that does them little good if the mailed ballot itself does not ever arrive, or if it arrives too late for a voter to return it in time to be counted. Defendants justify the law by contending that faxing or emailing ballots requires significant time and energy from municipal clerks. They also contend that there is a higher chance of human error because clerks have to re-create electronically returned ballots in paper form on election day, and that this process invades the voter’s privacy because those officials will see the voter’s selections. And a voter who receives an electronic copy of a ballot could forward that ballot to other voters, who might incorrectly believe that they can vote with it. According to defendants’ expert, Dr. Hood, these considerations supported the state’s decision to do away with faxing and emailing ballots to most absentee voters. DX001, at 19. As to the specific instances in which voters have had difficulty with receiving or sending absentee ballots by mail, defendants contend that voters can overcome these difficulties with planning, and they observe that electronic methods for sending ballots may not be any more reliable than using mail. Defendants’ justifications are not persuasive. Wisconsin already requires municipal clerks to send ballots by fax or email to military voters and to voters who are permanently 85 overseas, which undercuts most of defendants’ justifications. At trial, defendants principally relied on the testimony of two municipal clerks to defend this law. See Tr. 4p, at 141:12142:25; Tr. 7a, at 116:11-118:8. These clerks testified that electronic ballots can create a little more work before and on election day. Defendants did not present evidence of widespread opposition to sending ballots by fax or email. Indeed, other election officials could not see reasons for eliminating the practice, or testified that it did not create significant logistical problems. Tr. 2, at 332:23-333:4 (“It took a few minutes to compile the email.”), 333:15-17; PX435, at 48. From a practical perspective, the court simply does not credit the assertion that in the year 2016, printing a paper ballot and instructions, putting them into an envelope, and physically sending the envelope overseas is less burdensome on municipal clerks than compiling a PDF and sending an email. This is especially so because clerks are already sending ballots electronically to military and overseas electors. Defendants also overstate their concerns about privacy, security, and errors. A voter who chooses to submit an absentee ballot electronically is voluntarily giving up some of the privacy that a mailed ballot would have. That is the voter’s problem, not the state’s problem: a voter who is concerned about privacy can simply avoid voting by fax or email. As for defendants’ concern that voters may forward electronic copies of absentee ballots, they presented only one example of this occurring. There is no reason to think that it is a widespread problem. Even if it occurs regularly, a municipal clerk can correct the issue with an email to the voter who submitted a forwarded ballot. Finally, even crediting defendants’ assertion that there is a higher chance for human error when re-creating an electronically received ballot in paper form, that chance is minimal because two election officials perform 86 the task together. Defendants did not adduce evidence that mistakes ever actually happened, or that they happen with any frequency. If the challenges of sending and receiving electronic ballots are as severe as defendants make them out to be, then the state can make the practice optional instead of mandatory.22 But the state’s justifications for flatly prohibiting clerks from sending ballots by fax or email do not outweigh the moderate burdens that the challenged provision places on voters who are affected by it. The court concludes that the provision prohibiting municipal clerks from sending absentee ballots by fax or email violates the First and Fourteenth Amendments. 10. Limiting when clerks can return absentee ballots to voters Act 227, passed in 2012, prevents clerks from returning a received absentee ballot to a voter unless the ballot is damaged or has an incomplete certification. Plaintiffs contend that these provisions place undue burdens on voters with lower levels of educational attainment, who tend to be African Americans and Latinos. The court finds that the provisions governing when clerks can return absentee ballots to voters place only a slight burden on the right to vote. After Act 227, municipal clerks cannot return absentee ballots to voters to correct mistakes such as over-voting or improper marks. According to plaintiffs, minorities are more likely to make these kinds of mistakes because they have lower levels of educational attainment. PX036, at 9. Dr. Lichtman opined that “[t]his problem is especially acute for Wisconsin Hispanics. According to the US Census American Community Survey 2010, 3Year Estimates, 33.2 percent of Hispanics in Wisconsin speak English ‘less than very well.’” 22 Before 2011, the statute was permissive, not mandatory. 87 Id. at 48. The court does not give these opinions much value because Dr. Lichtman did not link his conclusion to the voting context. He did not identify what percentage of minority voters would have difficulty understanding a ballot, nor did he explain whether (and why) absentee ballots would be a type of printed document that minority voters would struggle to understand. Likewise, plaintiffs have not directed the court to any evidence demonstrating that comprehension problems with absentee ballots actually occur. See Dkt. 207, at 67. Defendants’ justification for this provision is straightforward and persuasive. Election officials do not open absentee ballots until election day, when they feed the ballots through counting machines. Thus, the only time that clerks would see the types of mistakes that plaintiffs identify is when they are actually preparing to feed the ballots through the machines. At that point, it is too late to return the ballot to the voter. In contrast, the errors for which clerks are now allowed to return absentee ballots are visible without opening the ballot envelope: “a spoiled or damaged absentee ballot,” Wis. Stat. § 6.86(5), and “an absentee ballot with an improperly completed certificate or with no certificate,” id. § 6.87(9). Beyond the procedural justification, defendants argue that permitting clerks to return ballots to correct “mistakes”—as plaintiffs want—leaves clerks without any real guidance. One clerk could determine that a voter made a mistake by not voting for each office on a ballot, while a different clerk could determine that the same voter apparently did not want to vote for each office. Preventing ambiguity and confusion serves the state’s interest in running efficient and orderly elections. The court concludes that the limits on when clerks can return absentee ballots to voters do not violate the First and Fourteenth Amendments. 88 11. The IDPP Plaintiffs contend that the IDPP impermissibly burdens the right to vote. They seek to invalidate the process not only for the petitioners who are currently trapped within it, but also for future petitioners who use the IDPP to obtain a free ID for voting purposes. The court finds that the IDPP imposes severe burdens on the right to vote. At least 60 qualified electors—those whose petitions were denied—were disenfranchised for the 2016 spring primary in Wisconsin. There were also 36 people in “suspend” status who had not been issued IDs. There is no evidence that any of these people were not qualified electors. And as defendants’ expert, Dr. Hood, acknowledged, there are “undoubtedly” people who are discouraged from even entering the process because they lack the documents or think that it is too cumbersome. Tr. 7p, at 199:11-200:8. Even petitioners who succeed in navigating the IDPP do so only after enduring severe burdens. Becky Beck, a CAFU research agent, indicated that once a petition gets to CAFU, it typically takes five separate contacts between the investigator and the petitioner to verify the petitioner’s identity, birthdate, and citizenship. Tr. 8p, at 159:12-16. CAFU’s Case Activity Reports document many instances in which petitioners are repeatedly sent to family members, hospitals, or schools to hunt for additional documentation, even when there is no doubt that the person is a qualified elector. Sometimes these petitioners succeed—but only after they have engaged in months of back-and-forth with CAFU—when the DMV finally determines, in its discretion, that the petitioner has made a strong enough case to warrant issuing an ID. Even when the effort is ultimately successful, the IDPP imposes burdens that far exceed those contemplated in Crawford and Frank. 89 Defendants invoke the same justifications that Crawford and Frank discuss. They contend that Wisconsin’s voter ID law (which includes the IDPP) deters fraud, promotes public confidence in elections, and promotes the orderly administration of elections. These interests justify a voter ID law in general, but they do not justify the severe burdens that the IDPP imposes. The Seventh Circuit has anticipated that such burdens could pose constitutional problems for Wisconsin’s voter ID law; it noted in Frank that: Milwaukee Branch of NAACP and the regulations leave much to the discretion of the employees at the Department of Motor Vehicles who decide whether a given person has an adequate claim for assistance or dispensing with the need for a birth certificate. Whether that discretion will be properly exercised is not part of the current record, however, and could be the subject of a separate suit if a problem can be demonstrated. 768 F.3d at 747 n.1. The evidence presented at trial confirms that the IDPP disenfranchises otherwise qualified voters. And even when confronted with lawsuits in two different federal courts, the state has utterly failed to devise a workable solution for getting these voters IDs. The state’s most recent emergency rule allows the petitioners who are currently in the IDPP to vote in the November 2016 election. But there is no plan in place for after the petitioners’ current receipts expire. Kicking the problem down the road does not alleviate the severe burdens that these petitioners must endure, nor does it prevent any future petitioners from suffering the same severe burdens. In short, many IDPP petitioners face insurmountable obstacles that serve no important interest because the government concedes that these petitioners are qualified electors. These justifications, such as they are, do not outweigh the burdens that the IDPP imposes. 90 The court concludes that the current version of the IDPP violates the First and Fourteenth Amendments. 12. Cumulative effect Plaintiffs contend that the cumulative effect of the challenged provisions in this case imposes an undue burden on the right to vote. According to plaintiffs, even if individual provisions comport with the First and Fourteenth Amendments, the court must still consider the overall effect of Wisconsin’s election system on voters, particularly on Democratic voters. To prove this aspect of their case, plaintiffs rely heavily on the “calculus of voting” theory that Dr. Burden explained in his expert report. PX037, at 4-5. Under this theory, a voter’s likelihood of voting is essentially the result of a formula that reflects a cost-benefit analysis. A person will vote if his or her probability of determining the outcome of the election, multiplied by the net psychological benefit of seeing his or her preferred candidate win, is greater than the “cost” of voting (i.e., the effort needed to become informed, and the time and resources needed to register to vote and cast a ballot). Id. Plaintiffs argue that Wisconsin has imposed a series of independently minor burdens that, collectively, increase the cost of voting enough to deter voters who tend to vote for Democrats. As explained above, plaintiffs did not present compelling statistical evidence of the deterrent effects that the challenged provisions have. But the nature of the challenged provisions, none of which facilitate voting or registration, makes it reasonable to infer that there will be some such effect. And as the Seventh Circuit recognized in Frank, “any procedural step filters out some potential voters.” 768 F.3d at 749 (original emphasis). But a deterrent effect alone, especially one that is not reliably quantified, does not render the cumulative effect somehow unconstitutional. 91 The Anderson-Burdick framework requires the court to evaluate “the precise interests put forward by the State as justifications for the burden imposed by its rule.” Burdick, 504 U.S. at 434. This requirement is difficult in the context of “cumulative effects” because the state can have different justifications for different rules, each with varying levels of persuasiveness. Plaintiffs do not propose a legal framework for evaluating a “cumulative effects” claim under Anderson-Burdick. But even looking broadly at the laws that they challenge in this case, the court’s analysis of the individual provisions already addresses the problematic aspects of Wisconsin’s election system. Take the challenged registration provisions: the court agrees that aspects of Wisconsin’s registration requirements burden the right to vote, particularly for voters who are more likely to move (which includes minority and younger voters, and thus, Democratic voters) and for voters who lack convenient access to documentary proof of residence (again, minority and younger voters, and thus, Democratic voters). But the state’s interests in preempting fraud, avoiding confusion, and ensuring that only qualified voters register to vote are compelling enough to justify at least some of the burdens that the challenged provisions collectively impose. Removing the restrictions on using dorm lists and reducing the durational residency requirement will ease the burdens of Wisconsin’s registration laws, at least to a degree that the state’s interests can justify. Likewise, the principal problem with Wisconsin’s in-person absentee system is that it addresses inequality across municipalities by suppressing voting in larger cities rather than by enabling increased voting in smaller cities. Invalidating that approach not only addresses the burdens on in-person absentee voting, but it also alleviates burdens in other aspects of Wisconsin’s election system. A voter who is intimidated by election observers or who is 92 concerned about long lines at the polls because there is no straight-ticket voting, for example, may be able to vote in-person absentee and avoid those concerns altogether. In short, although plaintiffs press a separate claim for the cumulative effects of the challenged provisions, the court concludes that they are entitled to no broader relief than the invalidation of the specific provisions that the court has identified as constitutionally infirm. A remedy directed at the diffuse cumulative effects of Wisconsin’s election regime would invite, essentially, a rewrite of the state’s election laws. That would be an unwarranted intervention by a federal court into an area reserved to the state legislature. F. Voting Rights Act claims Plaintiffs challenge the following provisions under § 2 of the Voting Rights Act: the reductions to in-person absentee voting; the one-location rule for in-person absentee voting; the elimination of corroboration; the requirement of documentary proof of residence; the elimination of statewide SRDs; the increased durational residency requirement; the zone for election observers; and the elimination of straight-ticket voting. Plaintiffs contend that these provisions disparately burden African Americans and Latinos. Section 2 of the Voting Rights Act prohibits states and political subdivisions from implementing any “voting qualification or prerequisite to voting or standard, practice, or procedure . . . in a manner which results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color.” 52 U.S.C. § 10301(a). Plaintiffs can establish a violation of § 2 by showing that, based on the totality of the circumstances, Wisconsin’s election process is “not equally open to participation by members of a class of [protected] citizens . . . in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice.” 93 Id. § 10301(b). Plaintiffs do not need to adduce proof of discriminatory intent to prevail on their Voting Rights Act claims. Chisom v. Roemer, 501 U.S. 380, 394-95 (1991). Most case law applying § 2 of the Voting Rights Act pertains to so-called “vote dilution” claims, which generally involve gerrymandering. Plaintiffs in this case bring claims over voting and registration requirements, which are “vote denial” claims for which Voting Rights Act law is less developed. In Frank, the Seventh Circuit endorsed a two-step inquiry for reviewing vote-denial challenges to voting qualifications under the Voting Rights Act: First, the challenged standard, practice, or procedure must impose a discriminatory burden on members of a protected class, meaning that members of the protected class have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice. Second, that burden must in part be caused by or linked to social and historical conditions that have or currently produce discrimination against members of the protected class. 768 F.3d at 754-55 (citations and internal quotation marks omitted). But the Seventh Circuit also cautioned that “§ 2(a) does not condemn a voting practice just because it has a disparate effect on minorities.” Id. at 753. “It is better to understand § 2(b) as an equaltreatment requirement (which is how it reads) than as an equal-outcome command.” Id. at 754. The court must therefore analyze whether plaintiffs have proven that: (1) the challenged provisions impose disparate burdens on African Americans and Latinos; and (2) under the totality of the circumstances, these burdens are linked to the state’s historical conditions of discrimination. 1. Disparate burdens Two threshold issues affect how the court evaluates plaintiffs’ evidence of disparate burdens. First, defendants contend that the Voting Rights Act requires plaintiffs to couch 94 their evidence in terms of a departure from an “objective benchmark,” rather than a departure from what Wisconsin’s laws used to be. Dkt. 206, at 114. The Supreme Court has indicated that a different baseline is part of what distinguishes § 2 claims from § 5 claims: In § 5 preclearance proceedings—which uniquely deal only and specifically with changes in voting procedures—the baseline is the status quo that is proposed to be changed: If the change “abridges the right to vote” relative to the status quo, preclearance is denied, and the status quo (however discriminatory it may be) remains in effect. In § 2 or Fifteenth Amendment proceedings, by contrast, which involve not only changes but (much more commonly) the status quo itself, the comparison must be made with a hypothetical alternative: If the status quo “results in [an] abridgement of the right to vote” or “abridge[s] [the right to vote]” relative to what the right to vote ought to be, the status quo itself must be changed. Reno v. Bossier Par. Sch. Bd., 528 U.S. 320, 334 (2000) (original emphasis). But Reno and the other cases on which defendants rely are vote dilution cases; this is a vote denial case. The few other federal courts that have considered how to evaluate burdens in vote denial cases have determined that this distinction is important. Relying on the text of the Voting Rights Act, the Southern District of Ohio recently concluded that “the relevant benchmark is inherently built into § 2 claims and is whether members of the minority have less opportunity than other members of the electorate to participate in the political process and elect representatives of their choice.” Ohio Org. Collaborative, 2016 WL 3248030, at *39; see also Ohio State Conference of N.A.A.C.P. v. Husted, 768 F.3d 524, 556 (6th Cir. 2014) (“Section 2 vote denial claims inherently provide a clear, workable benchmark. . . . under the challenged law or practice, how do minorities fare in their ability ‘to participate in the political process’ as compared to other groups of voters?” (original emphasis) (quoting 42 U.S.C. § 1973(b), which has been transferred to 52 U.S.C. § 10301)), vacated on other grounds, No. 14-3877, 2014 WL 10384647 (6th Cir. Oct. 1, 2014). The reasoning in these cases is 95 persuasive, and the court rejects defendants’ argument that plaintiffs must identify an objective benchmark to prevail on their Voting Rights Act claims. Part of determining whether minority voters have less opportunity to participate than other members of the electorate may involve comparing the challenged provisions with the laws that they replaced. See League of Women Voters of N.C. v. North Carolina, 769 F.3d 224, 241-42 (4th Cir. 2014), cert. denied, 135 S. Ct. 1735 (2015); Ohio Org. Collaborative, 2016 WL 3248030, at *40 (“[A]n analysis of whether a change in law results in a decreased opportunity of minorities to vote as compared to other voters is exactly the type of analysis required by § 2 claims.”). But that is not to say that a given provision would violate the Voting Rights Act just because it leaves minority voters worse off than a prior law. The appropriate inquiry at this first step is whether the challenged provision burdens minority voters more than other voters. See Frank, 768 F.3d at 753. The second threshold issue concerns the type of evidence that the parties have presented to prove (or disprove) that African Americans and Latinos have suffered disparate burdens under the challenged provisions. Experts on both sides have presented extensive statistical evidence derived from election turnout data in Wisconsin over time. Given the information available about Wisconsin’s elections, turnout rates may be the best that the parties can offer. But raw turnout statistics reveal very little about the disparate burdens that a state’s election system imposes. For example, defendants tout the high turnout numbers for the April 2016 election—the first statewide election in which the voter ID law and other challenged provisions were in effect—as evidence that minorities are not suffering disparate burdens under Wisconsin’s election laws. Tr. 1a, at 60:8-17. But turnout in a given election depends on many factors, ranging from which offices are on the ballot to the amount of 96 money spent on campaigning and the contentiousness of the races. The April 2016 Wisconsin involved unusually sharply contested primaries on both sides, which undoubtedly contributed to the higher-than-average turnout for an April election. Tr. 2, at 42:10-43:9. One cannot infer from the high overall turnout that Wisconsin’s election laws have no impact, or that they have no differential impact on minorities. That is not to say that turnout statistics are utterly useless. Plaintiffs’ expert, Dr. Mayer, used the statewide voter database, correlated to a separate database of demographic and political information, to track several cohorts of voters across the 2010 and 2014 elections (i.e., before and after some of the challenged provisions went into effect). Both sides’ experts agreed that comparing midterm elections, rather than presidential elections, made sense, because Barack Obama’s presence on the ballot in 2008 and 2012 would likely skew minority turnout. And, although the usual constellation of factors affected voting in 2010 and 2014, a change in election law regime was one significant difference between those elections, and no one was aware of any other major factor likely to affect turnout. Dr. Mayer also opined that, based on survey research, in 2014 most voters believed that the voter ID law was in effect, even though it was actually still enjoined. Thus, Dr. Mayer was of the view that the 2014 election would be a good test of the impact of the laws challenged in this case. Dr. Mayer used statistical regression analysis to isolate some of the variables that contribute to a voter’s likelihood of voting. Based on this analysis, Dr. Mayer concluded that African Americans, Latinos, and those who lived in student wards, were slightly less likely to vote in the 2014 election than the average voter was. PX043, at 14 (updated Table 8). By contrast, in the 2010 election, African Americans and those in student wards were actually 97 more likely to have voted. For Latinos, the difference between 2010 and 2014 was small (though slightly in the opposite direction; they were slightly less likely to vote in 2010). Plaintiffs contend that Dr. Mayer’s analysis shows that they challenged the provisions decreased likelihood that minorities will vote. These conclusions are in line with other national studies, which conclude that voter ID laws tend to suppress minority turnout at elections. See PX072. Defendants’ expert, Nolan McCarty, PhD, criticized Dr. Mayer’s conclusions because Dr. Mayer does not account for “roll-off” in the statewide voter database. That database provides a “snapshot” in that it includes voting records only for those voters who are registered as of the date the report of the database is generated, which in Dr. Mayer’s case was September 24, 2015. Thus the September 24, 2015 database does not include the voting records of any voter who was not registered as of that date, even though that voter might have been registered for the 2010 or 2014 elections. Dr. McCarty surmises that minority voters would have been more likely to rolloff, so that Dr. Mayer’s turnout rates for 2010 were too high, and thus the difference between those rates and the 2014 rates would be smaller. DX005, at 9. Dr. Mayer response is that despite the roll-off effect, his conclusions are sound, because he finds the effect even among the cohort of committed voters (because they stayed registered from 2010 to 2015 without rolling off the database). The court finds that, despite Dr. McCarty’s criticism, Dr. Mayer’s regression analysis supports the conclusion that the probability of an African American voting, relative to an average voter, was less in 2014 than it was in 2010. The court finds that Dr. Mayer’s conclusions about those who live in student wards are not informative, because his definition of those who live in student wards does not include only students. The bottom line is that Dr. Mayer’s analysis lends some support to the 98 plaintiffs’ claim that the challenged provisions tend to reduce African American voting by some modest amount. But nothing presented by either side demonstrated that the challenged laws had a striking impact on turnout overall or among any class of voters. And even with the support of other empirical evidence, Dr. Mayer’s conclusions, without more, are not enough to carry the day for plaintiffs. “It is better to understand § 2(b) as an equal-treatment requirement (which is how it reads) than as an equal-outcome command.” Frank, 768 F.3d at 754. At the end of the day, turnout statistics report outcomes, not the burdens of the election regulations that might have influenced those outcomes. Thus, the court must look for specific evidence demonstrating that the challenged provisions fall disparately on minorities. a. Registration provisions Plaintiffs challenge three registration-related provisions under the Voting Rights Act: proof of residence, elimination of corroboration, and elimination of statewide SRDs. Plaintiffs contend that these provisions impose disparate burdens on minority voters, who are more likely to move than white voters are. The court accepts plaintiffs’ expert evidence that minority populations are more transient. PX036, at 47. If those populations register at the same rate that white populations do, then they would need to complete registration more often. For minority voters who do not have convenient access to proof of residence, this requirement could be disparately burdensome, as could the elimination of corroboration. Wisconsin’s registration requirements apply to all voters, regardless of race. The fact that voters must register after they move does not itself impose a disparate burden. Instead, plaintiffs must demonstrate that it is categorically more difficult for African American or Latino voters to comply with the registration requirements, and that registering more often 99 therefore forces these populations to confront those difficulties more often. Plaintiffs have failed to make this showing. Even acknowledging that minorities are more likely to lack driver licenses or stateissued IDs, those are only 2 of the 12 options for proving residence that Wis. Stat. § 6.34(3)(a) authorizes. Dr. Lichtman indicates that minorities are more likely to be unemployed, id. at 7-8, which could mean that they would lack access to paychecks. But that still leaves residential leases, utility bills, bank statements, and documents issued by any unit of government. Indeed, as discussed above, municipal clerks have devised a strategy for sending letters to voters and then letting them use those letters to register. See, e.g., Tr. 1p, at 163-65; Tr. 2, at 301-02. Plaintiffs therefore cannot demonstrate that the documentary proof of residence requirement burdens minorities for purposes of § 2. Cf. Frank, 768 F.3d at 75253 (“[P]ersons who rely on the waiver procedure still must apply for it, which means that on average black and Latino residents must file more paperwork than white residents. Although these findings document a disparate outcome, they do not show a ‘denial’ of anything by Wisconsin, as § 2(a) requires.”). As for corroboration, plaintiffs’ evidence of a disparate burden substantially consists of anecdotes and lay observations. See, e.g., Tr. 1p, at 78:7-20 (corroboration is useful to people who are transient or in poverty); Tr. 3a, at 88:15-20 (corroboration facilitates participation by homeless or marginally housed voters). This testimony does not establish a verifiable disparate effect. And although some voters have been unable to register at the polls because corroboration is no longer an option, plaintiffs do not identify a racial slant to this problem. In fact, Dr. Lichtman expressly acknowledged that statistics about the use of corroboration 100 by race are not available. PX036, at 40. This leaves plaintiffs unable to prove that the elimination of corroboration disparately prevents minorities from registering to vote. In the abstract, African Americans and Latinos could have more difficulties presenting documentary proof of residence, particularly without corroboration. But plaintiffs have not actually proven that the challenged burdens disparately burden minorities. There is no persuasive evidence that minorities who want to register are systematically unable to comply with the requirement that they present proof of residence. The challenged provision violates the Voting Rights Act only if it gives “members of the protected class . . . less opportunity than other members of the electorate to participate in the political process.” Frank, 768 F.3d at 755 (citations and internal quotation marks omitted). Given the number of documents that voters can use to prove their residence, African American and Latino voters do not have “less opportunity” to participate in elections just because they are less likely to be able to use certain types of documents. Cf. Ohio Org. Collaborative, 2016 WL 3248030, at *40 (prohibiting officials from sending unsolicited applications for absentee ballots does not create a burden for § 2 purposes). Plaintiffs also argue that minority voters are more likely to register through SRDs at voter-registration drives than white voters are. But plaintiffs’ only citation for this proposition is a website. See Dkt. 207, at 204. Plaintiffs did not introduce the website as evidence at trial, and they do not direct the court to other evidence admitted at trial that supports this contention. The court therefore concludes that plaintiffs have failed to prove that the elimination of statewide SRDs has had a disparate effect on minorities. 101 The court concludes that the challenged provisions requiring documentary proof of residence, eliminating corroboration, and eliminating statewide SRDs do not disparately burden African Americans or Latinos. b. Durational residency provision In the context of plaintiffs’ constitutional challenge, the court concluded that the increased durational residency requirement imposes disparate burdens on African Americans and Latinos. For substantially the same reasons, the court concludes that this provision also disparately burdens minorities for purposes of the plaintiffs’ Voting Rights Act claims. Wisconsin’s minority populations are much more transient than its white population is, in terms of both moving into the state and moving within the state. PX036, at 47. Unlike the methods for proving residence, there is no flexibility in the durational residency requirement: a voter either satisfies the requirement or does not satisfy it. Voters who have not been in a municipality for at least 28 days must either return to their former municipalities (if they moved within Wisconsin) or be disenfranchised. Because African Americans and Latinos are also more likely to lack access to transportation and to have less flexible work schedules, traveling to another municipality is not always feasible. On top of these burdens, voters who first have to register in their former municipalities must complete the awkward process of certifying that they have “resided at the [former] residential address for at least 28 consecutive days immediately preceding this election, with no present intent to move.” DX101, at 1. The court concludes that the durational residency provision disparately burdens African Americans and Latinos. 102 c. In-person absentee voting provisions In the context of plaintiffs’ constitutional challenge, the court concluded that Wisconsin’s in-person absentee voting provisions burden the right to vote, particularly for minority populations in larger municipalities. For substantially the same reasons, the court concludes that these provisions also disparately burden minorities for purposes of plaintiffs’ Voting Rights Act claims. Wisconsin’s rules for in-person absentee voting all but guarantee that voters will have different experiences with in-person absentee voting depending on where they live: voters in large cities will have to crowd into one location to cast a ballot, while voters in smaller municipalities will breeze through the process. And because most of Wisconsin’s African American population lives in Milwaukee, the state’s largest city, the in-person absentee voting provisions necessarily produce racially disparate burdens. Moreover, plaintiffs have demonstrated that minorities actually used the extended hours for in-person absentee voting that were available to them under the old laws. PX036, at 43. The court concludes that the in-person absentee voting provisions disparately burden African Americans and Latinos. d. Election observer and straight-ticket voting provisions Plaintiffs contend that African Americans and Latinos are disparately affected by the state’s rules governing where election observers can stand at polling places and by the state’s elimination of straight-ticket voting. Problems with election observers are more prevalent in high-minority areas like Milwaukee and Racine. But, as with plaintiffs’ constitutional challenges to this provision, the problem for plaintiffs’ Voting Rights Act claims is that municipal clerks and chief election 103 inspectors decide where observers stand, not the state. The individual decisions that election officials make may lead to increased harassment at certain polling places. But that is not the same as saying that the state has imposed a disparate burden on minorities just by defining a range in which to position observers. Plaintiffs rely exclusively on anecdotal evidence to prove that observers intimidate or harass African Americans and Latino voters more often than white voters. This evidence is insufficient to prove a violation of the Voting Rights Act, and most of it is not directly relevant. Plaintiffs have not presented evidence—expert or otherwise—that minorities disparately suffer burdens when election observers stand close to them, or that the state’s zone for election observers leads election officials to place observers closer to voters in minority-heavy municipalities. Indeed, plaintiffs’ anecdotal evidence does not address the distances at which observers have caused problems, except to suggest that many observers were closer than three feet. That is not a result of Act 177—the state prohibited election officials from allowing observers to be closer than three feet. Thus, plaintiffs cannot attribute these problems to the state for purposes of proving a disparate burden. This leaves plaintiffs’ evidence that problems are more prevalent in Milwaukee and Racine. These problems occurred under the GAB’s rule, not under the statute that replaced it, which undermines plaintiffs’ assertion that Act 177 disparately burdens minorities. But even inferring that problems are more common in these municipalities under the new rule, the burden that minorities experience still comes from election officials not using the authority that the state has given them to control election observers. Plaintiffs have not proven that the state has imposed a disparate burden on African Americans or Latinos by 104 giving election officials discretion to designate zones for election observers that are appropriate for their polling locations. As for the elimination of straight-ticket voting, the court has already found that this provision imposes only slight burdens on the right to vote. For substantially similar reasons, the court concludes that the provision does not create a disparate burden for purposes of plaintiffs’ Voting Rights Act claims. Again, plaintiffs’ evidence is entirely anecdotal and mainly establishes only that African Americans and Latinos would prefer to use straight-ticket voting. The elimination of straight-ticket voting applies to all voters, regardless of race. Plaintiffs have failed to prove that this provision gives minorities less opportunity to vote than other voters. The court concludes that the challenged provisions governing election observers and straight-ticket voting do not disparately burden African Americans or Latinos. e. The IDPP As explained above, the IDPP imposes a discriminatory burden on racial minority groups, meaning that their members have less opportunity than others do to participate in the political process. Plaintiffs have made a more than ample showing on this element. The court concludes that the IDPP disparately burdens African Americans and Latinos. 2. Caused by or linked to social and historical conditions The second step in analyzing a claim under the Voting Rights Act is to consider whether a discriminatory burden is “in part . . . caused by or linked to social and historical conditions that have or currently produce discrimination against members of the protected class.” Frank, 768 F.3d at 755. Having concluded that Wisconsin’s durational residency 105 requirement, provisions for in-person absentee voting, and IDPP disparately burden African Americans and Latinos, the court now considers whether those burdens are linked to social and historical conditions of discrimination. Plaintiffs contend that the court should apply the so-called Gingles factors to analyze their Voting Rights Act claims. The Supreme Court has endorsed these factors, at least in the context of vote dilution cases. See Thornburg v. Gingles, 478 U.S. 30, 44-45 (1986). But the Seventh Circuit has found them to be “unhelpful in voter-qualification cases,” Frank, 768 F.3d at 754, and so the court will not organize its analysis by factor. Nevertheless, the Voting Rights Act requires courts to examine “the totality of circumstances,” 52 U.S.C. § 10301(b), which essentially comprises the same inquiries that the Gingles factors address. Thus, plaintiffs’ evidence about Wisconsin’s history of discrimination and about the effects of past discrimination that minority groups suffer is relevant to their Voting Rights Act claims. Wisconsin has a relatively scant history of state-sanctioned discrimination. When Wisconsin became a state in 1848, its constitution did not extend the right to vote to African Americans; they obtained that right after the measure was passed at a statewide election in 1849. But the effect of the election remained in doubt until 1866, when the Wisconsin Supreme Court clarified that African Americans had the right to vote. See generally Gillespie v. Palmer, 20 Wis. 544 (1866). Other statewide policies (or lack thereof) have disparately affected minorities to some degree, even if they were not facially discriminatory. For example, from 1913 to 2006, only municipalities with more than 5,000 residents had to register voters. In other municipalities, voters did not have to register. According to Dr. Burden, the result of this practice was that “98% of blacks and 91% of Latinos lived in municipalities where registration was required. In 106 contrast, only 68% of whites lived in these municipalities.” PX037, at 11. Thus, until 2006, minorities in Wisconsin disproportionately faced more impediments to voting than white citizens faced. Few municipalities outside of Milwaukee provide election-related materials in languages other than English, despite the fact that the GAB makes these forms available for clerks to use, and no other municipality provides ballots in Spanish. Id. Given the significant percentages of Spanish-speaking voters in municipalities across the state, id.; PX036, at 48, Wisconsin’s failure to address the issue is significant. Plaintiffs’ other evidence of historical conditions of discrimination concerns Milwaukee. This makes sense, given that Milwaukee is home to most of the state’s minority population. Along with other large cities in the state, Milwaukee is where the disparate burdens that the challenged provisions impose are most prevalent. But under the Voting Rights Act, “units of government are responsible for their own discrimination but not for rectifying the effects of other persons’ discrimination.” Frank, 768 F.3d at 753. Thus, defendants have argued in this case that Milwaukee’s history of discrimination, which is technically not the state’s own discrimination, cannot give rise to liability under the Voting Rights Act. Drawing such a rigid distinction for purposes of plaintiffs’ Voting Rights Act claims would undermine the purposes of the law. See Chisom, 501 U.S. at 403 (“Congress enacted the Voting Rights Act of 1965 for the broad remedial purpose of ridding the country of racial discrimination in voting. . . . [T]he Act should be interpreted in a manner that provides the broadest possible scope in combating racial discrimination.” (citations, internal quotation marks, and alterations omitted)). But even assuming that the Voting Rights Act does not 107 impose liability on the state for a municipality’s discrimination—a questionable assumption—the act certainly prevents a state from enacting laws that interact with a municipality’s history of discrimination to impose disparate burdens. See Frank, 768 F.3d at 754 (“We are not saying that, as long as blacks register and vote more frequently than whites, a state is entitled to make changes for the purpose of curtailing black voting. Far from it; that would clearly violate § 2 [of the Voting Rights Act].”). Beginning with the in-person absentee provisions, there is evidence that the state legislature passed these laws, at least in part, to specifically address what it perceived to be a problem with larger municipalities, like Milwaukee. Legislators were concerned that these municipalities offered residents more opportunities to vote than smaller municipalities offered. For example, during a floor session in the state senate, proponents of limiting the window for in-person absentee voting specifically referred to nipping Milwaukee and Madison’s practices “before too many other cities get on board.” PX022, at 6. Even if the state was not directly responsible for creating the socioeconomic disparities that exist in Milwaukee and other larger cities, the in-person absentee provisions impose burdens because of those disparities. For these reasons, the court concludes that evidence of discrimination in Milwaukee is relevant to the causation element of plaintiffs’ Voting Rights Act claims. During the 1960s and 1970s, Milwaukee experienced considerable white flight. Although the city’s Common Council passed an open housing law, discriminatory housing practices continued to limit housing choices for African Americans, confining them to the inner city. PX037, at 12. Zoning regulations in the municipalities surrounding Milwaukee further reinforced the segregation. As a result, two-thirds of Wisconsin’s African American 108 residents now live in Milwaukee, which remains one of the most segregated cities in the country. Id. at 13. Coupled with segregated housing practices, Milwaukee has also had a difficult history with discrimination in education. In 1976—more than 20 years after Brown v. Board of Education—a federal judge concluded that Milwaukee’s schools were illegally segregated. Amos v. Bd. of Sch. Dirs. of Milwaukee, 408 F. Supp. 765 (E.D. Wis.), aff’d sub nom., Armstrong v. Brennan, 539 F.2d 625 (7th Cir. 1976), vacated, 433 U.S. 672 (1977). The case settled after going to the Supreme Court. But the results of educational inequality have persisted. In 2015, high school graduation rates in Wisconsin were 66 percent for blacks, 78 percent for Latinos, and 93 percent for whites.23 PX037, at 16. Most of the rest of plaintiffs’ expert evidence does not link to the disparate burdens that the in-person absentee provisions create. For example, Dr. Burden catalogs other instances of racial disparities in incarceration rates, income, and health. Id. at 15-18. Although this evidence is credible, it is only tangentially relevant to plaintiffs’ Voting Rights Act claims. Likewise, Dr. Burden’s analysis of other Gingles factors (i.e., racially polarized voting, race-based appeals in political campaigns, minority members elected to public office) does not bear directly on the disparate burdens that the court has found. Disparities in housing, education, and employment, have left minority groups condensed into high-density urban areas, which makes them particularly vulnerable to Wisconsin’s rules for in-person absentee voting. With only one location for in-person absentee voting, voters must travel farther than they would otherwise have to travel if municipalities could establish more locations. And basic math confirms that one location in a 23 Although these are statewide statistics, the problem is likely just as prevalent in Milwaukee. 109 larger municipality will have to contend with a larger volume of voters than one location in a smaller municipality will have to confront. Lower levels of educational attainment and employment decrease the flexibility that minority populations will have to spend time waiting in line to vote in-person absentee, which makes the reduced hours problematic as well. The court therefore finds that the burdens that Wisconsin’s in-person absentee provisions impose are linked to historical conditions of discrimination. These provisions are invalid under the Voting Rights Act. As for durational residency, African Americans and Latinos will have to deal with this requirement more often than white voters will because they move more often. These populations are also more likely to lack access to transportation, meaning that if they do not satisfy the durational residency requirement, they will be less able to travel back to vote in their former municipalities. But plaintiffs have not persuasively explained how these burdens are linked to the historical conditions of discrimination described above. “Section 2(a) forbids discrimination by ‘race or color’ but does not require states to overcome societal effects of private discrimination that affect the income or wealth of potential voters.” Frank, 768 F.3d at 753. The court therefore finds that the burdens that Wisconsin’s durational residency requirement imposes are not linked the historical conditions of discrimination. These provisions do not violate the Voting Rights Act. Finally, based on the evidence adduced at trial, the court cannot conclude that the burdens that the IDPP imposes are linked to historical conditions of discrimination in Wisconsin. Most of the problems that petitioners have had with getting through the IDPP relate to their inability to provide vital records to the DMV or to CAFU. But those failures tend to result from historical conditions of discrimination in the petitioner’s home state or 110 country. Under Frank, it is not clear that the Voting Rights Act authorizes the court to hold Wisconsin accountable for these conditions. See 768 F.3d at 753 (“The judge did not conclude that the state of Wisconsin has discriminated in any of these respects. That’s important, because units of government are responsible for their own discrimination but not for rectifying the effects of other persons’ discrimination.”). It would be up to the Seventh Circuit, not this court, to clarify the scope of the inquiry under § 2. Plaintiffs contend that this is an excessively narrow reading of the Voting Rights Act, because it would allow Wisconsin to ignore rank discrimination by other states. They may be right, but the result appears to follow from Frank. Because the IDPP is manifestly unconstitutional under the Anderson-Burdick framework, the court will invalidate the IDPP regardless of its status under the Voting Rights Act. G. Fourteenth Amendment claims for disparate treatment of voters Plaintiffs initially challenged three of the provisions at issue under the Fourteenth Amendment, alleging that the legislature lacked a rational basis for: (1) implementing a 28day durational residency requirement; (2) eliminating straight-ticket voting; and (3) excluding technical college, out-of-state, and other expired IDs as qualifying forms of voter ID. Dkt. 19, ¶¶ 164-69. The court dismissed the claims concerning Wisconsin’s durational residency requirement and straight-ticket voting. Dkt. 66, at 5-9. At summary judgment, plaintiffs dropped their challenge to excluding technical college IDs, and the court granted summary judgment to defendants on most of the rest of plaintiffs’ remaining rational basis claim. Dkt. 185, at 20-24. The court denied defendants’ motion for summary judgment with regard to plaintiffs’ challenge that the state lacked a rational basis for excluding expired college or university IDs from the list of qualifying forms of voter ID. 111 In their post-trial brief, plaintiffs purport to “continue to challenge the rational basis of excluding three forms of ID: 1) out-of-state driver’s licenses, 2) driving receipts issued under Wis. Stat. § 343.11, and 3) state ID card receipts.” Dkt. 207, at 128. Plaintiffs are free to pursue these issues on appeal, but the court has already entered summary judgment for defendants on these aspects of plaintiffs’ rational basis claims. Plaintiffs also note that at summary judgment, the court “ruled that excluding expired college or university IDs lacked a rational basis.” Id. at 128 n.32. That is incorrect. In denying defendants’ motion, the court did not affirmatively conclude that the state lacked a rational basis for excluding expired college or university IDs. As the pertinent section of the summary judgment opinion stated: “[a]t this point, defendants have failed to identify a rational basis for the legislature’s decision to exclude expired student IDs. The court will deny this aspect of defendants’ motion for summary judgment.” Dkt. 185, at 24. The court essentially concluded that defendants’ proffered justifications for excluding expired student IDs were insufficient, and that defendants would have to do better at trial if they wanted to overcome plaintiffs’ rational basis challenge. Ultimately, plaintiffs’ misreading of the summary judgment decision is immaterial because rational basis review focuses on the state’s justification for its actions, rather than on plaintiffs’ disagreement with those actions. “[A] classification neither involving fundamental rights nor proceeding along suspect lines is accorded a strong presumption of validity.” Heller v. Doe, 509 U.S. 312, 319 (1993). The court will uphold the state’s decision to exclude expired college or university IDs if defendants identify “a rational relationship between the disparity of treatment and some legitimate governmental purpose.” Id. at 320. Defendants did not need to produce evidence at trial to support the rationality of the state’s decision, nor 112 are they limited to the justifications that the legislature had in mind at the time that it passed the challenged provisions—any rational justification for the laws will overcome an equal protection challenge. Id. at 320-21. The state’s approach to college and university IDs is somewhat inconsistent. The state purports to have given students the flexibility and convenience to choose how to verify their identities at the polls. In addition to the other forms of acceptable ID that are available to citizens generally, students have the unique option of using the IDs that they receive from their schools. But that option is not as convenient as it appears. College or university IDs are acceptable only if they expire within two years after issuance. Wis. Stat. § 5.02(6m)(f). The standard ID that the University of Wisconsin-Madison—the state’s flagship university— issues does not comply with this requirement. Tr. 1p, at 173:2-174:18; Tr. 3a, at 44:13-21. Instead, UW-Madison offers a second, voting-specific ID to its students who want to use university-issued IDs to vote. Tr. 3a, at 45:15-46:19. Thus, in practice, the option to use a college or university ID does not provide much flexibility or convenience. The state has also taken considerable pains to limit the use of college or university IDs to current students only. The three requirements in Wis. Stat. § 5.02(6m)(f) are redundant: (1) the ID card itself must be unexpired; (2) the card must have an expiration date that is no more than two years after its date of issuance; and (3) the voter must present proof of current enrollment. If each of these requirements provided some additional level of protection against former students using their IDs to vote, then those requirements might be rational. But as it stands, defendants have not explained why any requirement beyond proof of current enrollment is necessary to protect against fraudulent voting with a college or university ID. 113 Nevertheless, plaintiffs’ rational basis claim challenges only the requirement that the ID card be unexpired when a voter presents it at the polls. Defendants argue that it is rational to require voters to present unexpired college and university IDs because voters can use these IDs only in conjunction with proof of enrollment. See Wis. Stat. § 5.02(6m)(f). According to defendants, the state reasonably has presumed that anyone with an expired ID is probably no longer enrolled at the issuing college or university. Thus, it makes no sense to allow a voter to use an expired college or university ID because that voter will not be able to also provide proof of enrollment. This is a circular argument. Worse, it is the exact argument that defendants presented at summary judgment. The court concluded that this argument was not persuasive for two reasons: First, defendants apparently make no room for the possibility that a student could be enrolled at an institution but have an expired student ID. If incoming freshmen at four-year universities receive student IDs that expire two years after issuance, then any junior or senior who fails to obtain a new student ID would have to find a different way to prove his or her identity. Second, unlike receipts for driver licenses and ID cards, expired student IDs are not later replaced with entirely different documents. Defendants therefore cannot rely on the same arguments about simplifying elections by eliminating unnecessary duplicative forms of ID. Dkt. 185, at 24. Repetition has not made defendants’ argument any more persuasive. At a macro level, the state’s concern with ensuring that only current students vote with student IDs may be rational. But Wis. Stat. § 5.02(6m)(f) adequately addresses that concern by requiring a voter to present proof of enrollment with the student ID. Adding the requirement that a voter’s college or university ID be unexpired does not provide any additional protection against fraudulent voting. If anything, this measure prevents otherwise qualified voters from voting simply because they have not renewed their IDs since beginning 114 school. Thus, even under an exceedingly deferential rational basis review, the state has failed to justify its disparate treatment of voters with expired IDs. The court concludes that requiring unexpired college or university IDs violates the Fourteenth Amendment. To be clear, the court is not concluding that voters have carte blanche to use expired college or university IDs at the polls; they must still comply with the other requirements of Wis. Stat. § 5.02(6m)(f). Plaintiffs have not directed their rational basis challenge to the requirement that a voter with a college or university ID also present proof of enrollment at the issuing institution. Nor have plaintiffs challenged the rational basis for permitting only IDs that expire no more than two years after issuance.24 These requirements still apply. The only thing that will change is that the ID card that a college or university student actually presents at the polls can be expired. CONCLUSION AND REMEDIES The court has identified several constitutional and statutory violations, and the court will grant declaratory and injunctive relief accordingly. For the challenged provisions relating to in-person absentee voting, Wisconsin’s statutes establishing a one-location rule, Wis. Stat. § 6.855-.86, violate the First and Fourteenth Amendments and § 2 of the Voting Rights Act. Likewise, the sections of Act 146 amending Wis. Stat. §§ 6.86(1)(b) to limit the days and times for in-person absentee voting violate the Fifteenth Amendment. These provisions, along with the sections of Act 23 that 24 Without the requirement that a voter present an unexpired college or university ID, it seems unnecessary to regulate the ID’s expiration date. But that is outside the scope of plaintiffs’ challenge, and so the court will leave it to the state to determine whether this provision is still necessary. 115 limit the hours for in-person absentee voting, also violate § 2 of the Voting Rights Act and the First and Fourteenth Amendments, except with regard to preventing municipal clerks from holding hours for in-person absentee voting on the Monday before an election. For the challenged provisions relating to registering to vote, the sections of Act 23 amending Wis. Stat. § 6.34(3)(a)7. to require dorm lists to include proof of a student’s citizenship violate the First and Fourteenth Amendments. Likewise, the sections of Act 23 amending Wis. Stat. §§ 6.02, .10(3), and .15 to increase the durational residency requirement from 10 days to 28 days violate the First and Fourteenth Amendments. For the challenged provisions relating to election procedures, the sections of Act 75 amending Wis. Stat § 6.87(3)(d) to prohibit municipal clerks from emailing or faxing absentee ballots to voters violate the First and Fourteenth Amendments. For the challenged provisions relating to voter ID, the statutes and administrative rules that create and govern the IDPP that voters can use to obtain free IDs for purposes of voting violate the First and Fourteenth Amendments. Plaintiffs seek a permanent injunction. Dkt. 207, at 244. They must therefore demonstrate that: (1) they have succeeded on the merits; (2) no adequate remedy at law exists; (3) they will suffer irreparable harm without injunctive relief; (4) the irreparable harm suffered without injunctive relief outweighs the irreparable harm that Wisconsin will suffer if the injunction is granted; and (5) the injunction will not harm the public interest. Old Republic Ins. Co. v. Emp’rs Reinsurance Corp., 144 F.3d 1077, 1081 (7th Cir. 1998). Based on the court’s conclusion that several of the challenged provisions violate the Constitution or the Voting Rights Act, or both, the court finds that plaintiffs have made the requisite showing 116 and injunctive relief is appropriate. With the exception of the IDPP, the court will permanently enjoin defendants from enforcing the invalid provisions. The IDPP does not require wholesale invalidation. As described in the introduction to this opinion, another federal court has already issued a preliminary injunction against enforcing the IDPP. That injunction imposes an affidavit-based solution, essentially allowing voters to sign a form instead of presenting an ID at the polls. Plaintiffs have not asked for that type of relief here, and the court will not grant it. Nothing would prevent the state from complying with both Judge Adelman’s injunction and the one that this court will impose. This court will require that the IDPP be reformed to satisfy two criteria. First, Wisconsin cannot make it unreasonably difficult for voters to obtain a free ID. Once a petitioner has submitted materials sufficient to initiate the IDPP, the DMV must promptly issue a credential valid for voting, unless readily available information shows that the petitioner is not a qualified elector entitled to such a credential. Second, the state must inform the general public that those who enter the IDPP will promptly receive a credential valid for voting, unless readily available information shows that the petitioner is not a qualified elector entitled to such a credential. For further clarification: the credentials issued under this procedure need not be valid for any purpose other than voting; the court is not ordering the state to issue Wisconsin IDs to all those who enter the IDPP. But the credentials issued are not temporary: petitioners and the public must be informed that these credentials have a term equivalent to that of a driver license or Wisconsin ID, and that they will be valid for voting until they expire or are revoked for good cause. Good cause is shown if the petitioner is not a qualified elector; the failure to provide additional information or communication to the DMV is not good cause. The 117 receipts issued under the most recent Emergency Rule would meet these requirements, with the exception of the currently stated term of expiration. ORDER IT IS ORDERED that: 1. The IDPP as implemented is unconstitutional under the First and Fourteenth Amendments to the United States Constitution; 2. 2013 Wis. Act 146 is unconstitutional under the Fifteenth Amendment to the United States Constitution; 3. The restriction limiting municipalities to one location for in-person absentee voting is unconstitutional under the First and Fourteenth Amendments to the United States Constitution; 4. The state-imposed limits on the time for in-person absentee voting, with the exception of the prohibition applicable to the Monday before election day, are unconstitutional under the First and Fourteenth Amendments to the United States Constitution; 5. The requirement that “dorm lists” to be used as proof of residence include citizenship information is unconstitutional under the First and Fourteenth Amendments to the United States Constitution; 6. The increase of the durational residency requirement from 10 days to 28 days is unconstitutional under the First and Fourteenth Amendments to the United States Constitution; 7. The prohibition on distributing absentee ballots by fax or email is unconstitutional under the First and Fourteenth Amendments to the United States Constitution; 8. The prohibition on using expired, but otherwise qualifying, student IDs is unconstitutional under the First and Fourteenth Amendments to the United States Constitution; 9. Plaintiffs’ request for a permanent injunction is GRANTED, and defendants are permanently enjoined from enforcing any of the provisions held unlawful in sections 1 through 8 of this ORDER; 10. Defendants, and their officers, agents, servants, employees, attorneys, and all those acting in active concert or participation with them, or having actual or implicit knowledge of this order, are further ORDERED to: 118 a. Promptly issue a credential valid as a voting ID to any person who enters the IDPP or who has a petition pending; b. Provide that any such credential has a term of expiration equivalent to that of a Wisconsin driver license or photo ID and will not be cancelled without cause; c. Inform the general public that credentials valid for voting will be issued to persons who enter the IDPP; d. Further reform the IDPP so that qualified electors will receive a credential valid for voting without undue burden, consistent with this opinion; 11. Provisions 10.a. through 10.d. are to be effectuated within 30 days so that they will be in place and available for voters well before the November 8, 2016, election. 12. The court retains jurisdiction to oversee compliance with the injunction; 13. The court intends this ruling to be immediately appealable; for the avoidance of doubt, the court grants permission to any party to file an interlocutory appeal if this order is not final for appeal purposes. Entered July 29, 2016. BY THE COURT: /s/ ________________________________________ JAMES D. PETERSON District Judge 119