Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 1 of 78 IN THE UNITED STATES DISTRICT COURT FOR THE NORTHERN DISTRICT OF GEORGIA ATLANTA DIVISION Lucille Anderson, et al., Plaintiffs, Case No. 1:20-cv-03263 Michael L. Brown United States District Judge v. Brad Raffensperger, et al., Defendants. ________________________________/ OPINION & ORDER Plaintiffs (voters and political organizations) filed a complaint and motion for preliminary injunction asking this Court to step into Georgia’s November 2020 Presidential Election to dictate how various counties equip, staff, and operate their polling locations, how they train poll workers, and how they function on Election Day. (Dkts. 1; 92.) They say all of this is necessary because these counties (and really all of Georgia) have “faced some of the longest average wait times to vote in the entire county, often waiting hours to vote, with many deterred from voting and disenfranchised.” (Dkt. 1 ¶ 3.) They claim these problems reached a crescendo in the June 2020 Primary when “Georgia’s election system Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 2 of 78 experienced a complete meltdown.” (Dkt. 92-1 at 1.) Finally arguing “what’s past is prologue,” Plaintiffs say “nothing will change without judicial compulsion,” guaranteeing lines that threaten to disenfranchise voters in the November 2020 Election. (Dkts. 92-1 at 2; 159 at 10.)1 Plaintiffs’ motion for preliminary injunction lists the steps it asks the Court to take to manage the November Election. Defendants move to dismiss for lack of subject matter jurisdiction or, alternatively, for failure to state a claim. (Dkts. 105; 106.) The Court grants Defendants’ motions to dismiss, denies Plaintiffs’ preliminary injunction motion, and dismisses this case for lack of subject matter jurisdiction. I. Background Plaintiffs are three individuals (Lucille Anderson, Sara Alami, and Gianella Contreras Chavez) and two political organizations (DSCC and Democratic Party of Georgia, Inc.). Plaintiffs Anderson, Alami, and Chavez (together, “Individual Plaintiffs”) are registered voters in Fulton County. (Dkts. 1 ¶¶ 12–14; 93-30 ¶¶ 2–3; 93-31 ¶¶ 2–3; 93-32 ¶ 2.) They Plaintiffs quote The Tempest to suggest what has happened before will happen again, that is, the past will repeat itself. Shakespeare may have had a different meaning: that what has happened in the past sets the stage for what is to come, including the opportunity for either greatness or failure. 1 2 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 3 of 78 each waited in long lines to vote in the June 2020 Primary. (Dkts. 1 ¶¶ 12–14; 93-30 ¶¶ 4–7; 93-31 ¶¶ 5–6; 93-32 ¶¶ 3–7.)2 Plaintiff Anderson went to her polling location on her way to work but, finding long lines, could not vote. She returned that afternoon, waited in line more than an hour, and went home because she feared she might “pass out” from the heat. She returned about an hour later, but left without voting when she discovered the line was even longer than before. She planned to return at 9:00 p.m. but was too “exhausted from a full day of work and waiting to vote in the Georgia heat” to do so. (Dkts. 1 ¶ 12; 93-30 ¶¶ 4–7.) Plaintiff Alami arrived at her polling location 30 minutes before it opened but still waited in line for 6 hours before voting. (Dkts. 1 ¶ 13; 93-31 ¶¶ 5–6.) Plaintiff Chavez waited in line for more than 8 hours, before finally voting at 1:00 a.m. (Dkts. 1 ¶ 14; 93-32 ¶¶ 3–7.) Each Individual Plaintiff plans to vote in the November 2020 General Election but worries she will have to wait in long lines in order to do so. (Dkts. 1 ¶¶ 12–14; 93-30 ¶ 9; 93-31 ¶ 8; 93-32 ¶ 8.) Although Plaintiff Chavez now resides in Fulton County, she lived and voted in Cobb County during the June 2020 Primary. (Dkts. 1 ¶ 14; 93-32 ¶ 2.) 2 3 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 4 of 78 Plaintiff DSCC is the national senatorial committee of the Democratic Party. (Dkts. 1 ¶ 15; 93-63 ¶ 3.) Its mission is to elect candidates of the Democratic Party to the U.S. Senate, including from Georgia. (Id.) Plaintiff Democratic Party of Georgia, Inc. (“DPG”) is a state committee responsible for the day-to-day operation of the Democratic Party in Georgia. (Dkt. 1 ¶ 16); 52 U.S.C. § 30101(15). Its mission is to elect Democratic Party candidates across the state. (Dkts. 1 ¶ 16; 93-60 ¶ 2.) Plaintiffs DSCC and DPG (together, “Organizational Plaintiffs”) claim long voting lines harm their missions and require them to divert resources as a result. (Dkts. 1 ¶¶ 15–16; 93-60 ¶¶ 17–19; 93-63 ¶¶ 10–11.) Defendants are Georgia state and county officials sued in their official capacities. The state Defendants are the Georgia Secretary of State and members of the Georgia State Election Board (together, “State Defendants”). The county Defendants are members of nine County Boards of Registration and Elections (together, “County Defendants”). Those boards are in Fulton, DeKalb, Cobb, Gwinnett, Chatham, Clayton, Henry, Douglas, and Macon-Bibb Counties (together, “Nine Counties”). 4 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 5 of 78 Plaintiffs claim Defendants have repeatedly caused long voting lines in the Nine Counties by failing to provide (1) enough polling locations, (2) enough voting equipment, (3) enough poll workers, (4) adequate training to poll workers, (5) enough technicians to address equipment malfunctions and other technical issues, (6) enough time to set up polling locations, and (7) enough backup paper pollbooks and emergency paper ballots in case voting equipment breaks down. (Dkt. 1 ¶¶ 4, 20; id. at 79.) Plaintiffs claim these failures unduly burden the right to vote in violation of the First and Fourteenth Amendments (Count 1); result in “extraordinary voting restrictions that render the Voting System fundamentally unfair” in violation of the substantive component of the Due Process Clause (Count 2); and “place[] widely different burdens on voters across the State . . . . depending on the counties in which they live,” in violation of the Equal Protection Clause (Count 3). (Id. ¶¶ 197–214.) The complaint seeks declaratory and injunctive relief. (Id. at 79.) In September 2020, Plaintiffs moved for a preliminary injunction asking the Court to order Defendants to take the following actions for the November 2020 Election: 5 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 6 of 78 (1) Defendants must use a queue formula submitted by Plaintiffs’ expert to “allocate voting machines, poll pads, scanners, and technicians in a way reasonably calculated to minimize wait times” at each polling location within the Nine Counties. (Dkt. 119-3 at 4–6.) (2) Defendants must supply each polling location within the Nine Counties with (a) “emergency paper ballots equal to or exceeding 40% of the number of registered voters assigned to a polling place”; (b) “sufficient paper pollbook backups updated to show who has voted through Friday, October 30, 2020, in the event of electronic poll book malfunction”; and (c) “sufficient secure ballot boxes to hold emergency paper ballots in the event of scanner malfunction.” (Id. at 6–7.) (3) State Defendants must “enact a policy requiring and issue guidance to all county election officials instructing that poll workers utilize the emergency backup supplies whenever the last voter in line is expected to or does in fact wait 30 minutes or more to cast a ballot.” (Id. at 7.) (4) Defendants must “enact a policy requiring poll workers to consistently monitor and record wait times every 30 minutes at polling places.” (Id. at 7–8.) (5) Defendants must “ensure that poll workers at polling locations within [the Nine Counties] are adequately trained to operate all components of the voting system . . . . [and] to address common equipment failure issues.” (Id. at 8.) (6) Defendants must ensure, at the Secretary of State’s expense, that there are “sufficient technicians available for deployment at each polling place within [the Nine Counties] on less than 30 minutes notice.” (Id. at 9.) (7) The Secretary of State must “allocate and train at least one technician for each 10 polling locations within a county.” (Id.) 6 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 7 of 78 (8) Defendants must “enact policies requiring that the functionality of polling equipment at each polling location be adequately tested at each polling location within a reasonable time before the location opens.” (Id.) Plaintiffs say these actions are required to avoid long voting lines at the November 2020 Election because Defendants will otherwise make the same mistakes that have caused long lines in the past. (See, e.g., Dkt. 1 ¶¶ 6, 208.) Defendants moved to dismiss in September 2020. They say the Court lacks subject matter jurisdiction for several reasons, including that Plaintiffs do not have standing to pursue their claims. They also seek dismissal for failure to state a claim. Plaintiffs sought no discovery and, in early October 2020, the Court held a hearing on the parties’ motions. (Dkt. 124.) The Court expressed reservations about whether Plaintiffs had presented facts sufficient to establish the imminent injury required for standing. (See, e.g., Dkt. 159 at 5–6.) In an effort to afford Plaintiffs every opportunity to make that showing, the Court ordered County Defendants to provide Plaintiffs with additional information about their proposed allocation of voting equipment and voters for the November 2020 Election. (Dkt. 159 at 123– 127.) County Defendants have since provided Plaintiffs with that information, as well as information about the steps they have taken 7 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 8 of 78 following the June 2020 Primary to avoid the problems that plagued that election. After Plaintiffs filed an expert affidavit predicting “long” lines at several precincts, the Court held a second evidentiary hearing. (Dkts. 149-1; 169.) Despite this expert affidavit, the Court’s reservations about Plaintiffs’ standing have only increased. It is now clear that Plaintiffs lack standing to bring their claims because they have not shown they face a certainly impending injury. II. Plaintiffs Lack Standing Plaintiffs claim they will be injured by long voting lines in the Nine Counties during the November 2020 Election. Individual Plaintiffs say these lines will burden their right to vote. (Dkt. 111 at 3–4.) And Organizational Plaintiffs say the lines will require them to divert resources to help voters (including their own members) avoid disenfranchisement. (Id. at 8–9, 11.) Because Plaintiffs point to future injuries that depend on the existence of long lines at the November 2020 Election, Plaintiffs must show those lines are sufficiently likely to occur — or, in the language of the caselaw, that they are “certainly impending” 8 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 9 of 78 — in order to establish standing. Plaintiffs say the lines are “all but certain” to occur. (Dkt. 118 at 17.) The Court disagrees.3 A. General Standing Principles “Article III of the Constitution limits federal courts to adjudicating actual ‘cases’ and ‘controversies.’” A&M Gerber Chiropractic LLC v. GEICO Gen. Ins. Co., 925 F.3d 1205, 1210 (11th Cir. 2019). The doctrine of standing “constitutes the core of Article III’s case-or-controversy requirement.” Steel Co. v. Citizens for a Better Env’t, 523 U.S. 83, 103– 04 (1998); see A&M Gerber Chiropractic, 925 F.3d at 1210 (“Perhaps the most important of the Article III doctrines grounded in the case-orcontroversy requirement is that of standing.”). “Standing cannot be waived or conceded by the parties, and it may be raised (even by the court To the extent Organizational Plaintiffs are diverting resources because they believe long lines are inevitable, that is insufficient to establish standing — they must show the lines are actually likely to occur. See Clapper v. Amnesty Int’l USA, 568 U.S. 398, 402 (2013) (“[Plaintiffs] cannot manufacture standing by choosing to make expenditures based on hypothetical future harm that is not certainly impending.”); City of Los Angeles v. Lyons, 461 U.S. 95, 107 n.8 (1983) (“It is the reality of the threat of repeated injury that is relevant to the standing inquiry, not the plaintiff’s subjective apprehensions.”); Greater Birmingham Ministries v. State, 161 F. Supp. 3d 1104, 1114 (N.D. Ala. 2016) (“[S]tanding based on diverting resources to avoid the risk of hypothetical future harm is not a sufficient injury.”). 3 9 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 10 of 78 sua sponte) at any stage of the case.” A&M Gerber Chiropractic, 925 F.3d at 1210. Courts “have always insisted on strict compliance with this jurisdictional standing requirement.” Raines v. Byrd, 521 U.S. 811, 819 (1997). “[T]o satisfy Article III’s standing requirements, a plaintiff must show (1) it has suffered an injury in fact that is (a) concrete and particularized and (b) actual or imminent, not conjectural or hypothetical; (2) the injury is fairly traceable to the challenged action of the defendant; and (3) it is likely, as opposed to merely speculative, that the injury will be redressed by a favorable decision.” Friends of the Earth, Inc. v. Laidlaw Envtl. Servs. (TOC), Inc., 528 U.S. 167, 180–81 (2000). “The plaintiff bears the burden of establishing each element.” Cordoba v. DIRECTV, LLC, 942 F.3d 1259, 1268 (11th Cir. 2019). And plaintiff must meet that burden for “each claim against each Defendant.” Warren Tech., Inc. v. UL LLC, 2018 WL 10550930, at *5 (S.D. Fla. Oct. 31, 2018); see DaimlerChrysler Corp. v. Cuno, 547 U.S. 332, 352 (2006) (“[A] plaintiff must demonstrate standing for each claim he seeks to press.”). Courts “should not speculate concerning the existence of standing, nor should [they] imagine or piece together an injury sufficient to give 10 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 11 of 78 plaintiff standing when it has demonstrated none.” Bochese v. Town of Ponce Inlet, 405 F.3d 964, 976 (11th Cir. 2005). In other words, “[i]t is not enough that [plaintiff] sets forth facts from which [the court] could imagine an injury sufficient to satisfy Article III’s standing requirements.” Id. Instead, “plaintiff has the burden to clearly and specifically set forth facts sufficient to satisfy Art. III standing requirements.” Id.; see Spokeo, Inc. v. Robins, 136 S. Ct. 1540, 1547 (2016) (noting that, even “at the pleading stage, the plaintiff must clearly allege facts demonstrating each element” of standing). “If the plaintiff fails to meet its burden, this court lacks the power to create jurisdiction by embellishing a deficient allegation of injury.” Bochese, 405 F.3d at 976. “[A] motion to dismiss for lack of subject matter jurisdiction pursuant to Fed. R. Civ. P. 12(b)(1) can be based upon either a facial or factual challenge to the complaint.” McElmurray v. Consol. Gov’t of Augusta-Richmond Cty., 501 F.3d 1244, 1251 (11th Cir. 2007). “A facial attack on the complaint requires the court merely to look and see if the plaintiff has sufficiently alleged a basis of subject matter jurisdiction, and the allegations in his complaint are taken as true for the purposes of the 11 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 12 of 78 motion.” Id. “Factual attacks, on the other hand, challenge the existence of subject matter jurisdiction in fact, irrespective of the pleadings, and matters outside the pleadings, such as testimony and affidavits are considered.” Id. Defendants lodge a factual attack here. (See, e.g., Dkts. 105-1 at 3–4, 6 (citing evidence outside the complaint and arguing “Plaintiffs’ allegations do not enjoy the same degree of deference” because this is “a challenge to jurisdiction in a factual 12(b)(1) motion”); 151 at 2– 6 (citing extrinsic evidence to challenge standing); 170-2 at 8–12 (acknowledging the Court can consider extrinsic evidence to determine standing).) This means the “trial court is free to weigh the evidence and satisfy itself as to the existence of its power to hear the case.” Makro Capital of Am., Inc. v. UBS AG, 543 F.3d 1254, 1258 (11th Cir. 2008).4 The significance of the facial/factual distinction may be limited. Even in a facial attack, courts “are obliged to consider not only the pleadings, but to examine the record as a whole to determine whether [they] are empowered to adjudicate the matter at hand.” Elend v. Basham, 471 F.3d 1199, 1208 (11th Cir. 2006); see Corbett v. Transp. Sec. Admin., 930 F.3d 1225, 1235 (11th Cir. 2019) (“While we typically confine our standing analysis to the four corners of the complaint, we may look beyond it when we have before us facts in the record.”). There is also authority that, in a facial attack, a court is still “free to weigh the evidence and satisfy itself as to the existence of its power to hear the case.” Flat Creek Transportation, LLC v. Fed. Motor Carrier Safety Admin., 923 F.3d 1295, 1299 n.1 (11th Cir. 2019) (rejecting the argument that a facial attack precluded the court from weighing the evidence). 4 12 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 13 of 78 A plaintiff must, however, have “ample opportunity to present evidence bearing on the existence of jurisdiction.” Morrison v. Allstate Indem. Co., 228 F.3d 1255, 1273 (11th Cir. 2000). Given the eleventh hour at which Plaintiffs filed and briefed their motions, they have certainly had that here. The parties have submitted substantial evidence on the issue of standing, the Court allowed Plaintiffs to obtain additional standing information from Defendants (even though they never requested it), and the Court has held two hearings at which the parties were entitled to introduce evidence (including one in which Plaintiffs’ expert testified). No one has requested more discovery, another hearing, or additional time to present evidence. B. Injury in Fact Injury in fact is “the first and foremost of standing’s three elements.” Spokeo, 136 S. Ct. at 1547. “When a plaintiff seeks prospective relief to prevent a future injury, it must establish that the threatened injury is certainly impending.” Indep. Party of Fla. v. Sec’y, State of Fla., 967 F.3d 1277, 1280 (11th Cir. 2020); see Jacobson v. Fla. Sec’y of State, --- F.3d --- 2020 WL 5289377, at *4 (11th Cir. Sept. 3, 2020) (same); see also Clapper, 568 U.S. at 410 (“[A] threatened injury must be 13 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 14 of 78 certainly impending to constitute injury in fact.”). “[A]llegations of possible future injury are not sufficient.” Clapper, 568 U.S. at 409. Nor is a “realistic threat,” Summers v. Earth Island Inst., 555 U.S. 488, 499– 500 (2009), an “objectively reasonable likelihood” of harm, Clapper, 568 U.S. at 410, or “a ‘perhaps’ or ‘maybe’ chance” of injury, Bowen v. First Family Fin. Servs., Inc., 233 F.3d 1331, 1340 (11th Cir. 2000). The Supreme Court has said that literal certainty is not “uniformly require[d],” and that a “substantial risk” or “substantial likelihood” of harm may be enough “[i]n some instances.” Clapper, 568 U.S. at 414 n.5 (emphasis added); Bowen, 233 F.3d at 1340 (emphasis added). The required showing is ultimately “a matter of degree,” Thompson v. State Farm Fire & Cas. Co., 2016 WL 2930958, at *3 (M.D. Ga. May 19, 2016), and “[h]ow likely is enough is necessarily a qualitative judgment,” Fla. State Conference of N.A.A.C.P. v. Browning, 522 F.3d 1153, 1161 (11th Cir. 2008). C. Defendants’ Past Elections Certainly Impending Injury Do Not Establish a Plaintiffs’ prediction of long lines in November 2020 is based almost entirely on the existence of long lines in past elections. Although “past wrongs are evidence bearing on whether there is a real and immediate 14 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 15 of 78 threat of repeated injury,” O’Shea v. Littleton, 414 U.S. 488, 496 (1974), they are typically “insufficient alone,” Aransas Project v. Shaw, 775 F.3d 641, 648 (5th Cir. 2014); see Stanley v. Broward Cty. Sheriff, 773 F. App’x 1065, 1069 (11th Cir. 2019) (“Past exposure to illegal conduct does not in itself show a pending case or controversy regarding injunctive relief if unaccompanied by any continuing, present injury or real and immediate threat of repeated injury.”). That is, courts generally “must focus on what [defendant] has promised to do going forward,” even if defendant has a “long and sordid history” of violations. Hoffer v. Jones, 290 F. Supp. 3d 1292, 1300–01 (N.D. Fla. 2017); see Mancha v. Immigration & Customs Enf’t, 2007 WL 4287766, at *2 (N.D. Ga. Dec. 5, 2007) (“The focus of the inquiry is on prospective conduct. Therefore, it follows that past exposure to illegal conduct does not in itself show a present case or controversy regarding injunctive relief.”). Plaintiffs’ failure to focus on the future here does not foreclose standing as a matter of law. See Ciudadanos Unidos De San Juan v. Hidalgo Cty. Grand Jury Comm’rs, 622 F.2d 807, 820 (5th Cir. 1980) (finding “allegations of past illegal conduct . . . sufficient to give rise to a strong inference that the injury will be repeated in the future”). But it does make standing harder to establish. 15 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 16 of 78 1. Defendants’ Past Elections Were Different The predictive value of Georgia’s past elections is simply too limited to tell us (with the requisite certainty) what will happen in November. Georgia revamped its voting equipment and voting process in 2020, so elections before then reveal little about elections today. Before 2020, Georgia used Direct Recording Electronic voting machines that, in the words of one court, were “unreliable and grossly outdated.” Curling v. Raffensperger, 397 F. Supp. 3d 1334, 1403 (N.D. Ga. 2019). Georgia now uses a three-stage voting process that involves entirely different equipment: (1) voters check in using electronic poll pads; (2) they cast their ballots on machines known as Ballot Marking Devices (“BMDs”); and (3) they deposit their printed ballots through optical scanners. (Dkt. 93-62 at 10.) This new approach was introduced as part of Georgia House Bill 316,5 which made sweeping changes to the state’s voting system more generally. These changes dilute the link between Georgia’s past and future elections. Indeed, Plaintiffs’ own expert declined to use pre-2020 election data “in making assessments for the 2020 General Election Georgia Act No. 24, Georgia House Bill 316, amending Chapter 2 of Title 21 of the Official Code of Georgia Annotated. 5 16 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 17 of 78 because the voting machines and, as such, voting process for those elections are different from the process that will [be] in use in the 2020 General Election.” (Dkt. 93-62 at 1 n.1.) That leaves the June 2020 Primary. But Plaintiffs conceded at oral argument that, however bad that election was, it does not by itself show that the November 2020 Election will be similarly plagued by long lines. (See Dkt. 159 at 116.) The Court agrees. The June 2020 Primary was unique in ways that make it an unreliable guide to future elections. It was the first election in which Defendants used their new voting equipment and voting process. And it was the first election that Defendants administered during the global COVID-19 pandemic. (See, e.g., Dkts. 108-1 ¶ 4 (“[t]he June 9 primary was the most unique circumstances I have ever encountered in an election” due to the “new voting system” and the “global pandemic”); 108-2 ¶ 3 (same); 108-3 ¶ 6 (noting “unique challenges” based on “an unprecedented global pandemic” and “newer voting equipment”).) There is no doubt that these two features contributed substantially to long lines in June 2020. The introduction of new voting machines, for example, led to problems as simple as one county discovering on the eve 17 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 18 of 78 of Election Day that the new machines were so much bigger than the old ones that the county did not have enough trucks to deliver them to polling locations in time for opening. (Dkts. 142-1 ¶¶ 3–4; 159 at 84; see also Dkts. 108-1 ¶ 9; 108-2 ¶ 6.) To make matters worse, the pandemic limited the extent to which poll workers received training on these new machines. (See, e.g., Dkts. 128-1 ¶¶ 5 (“[O]pportunities for poll worker training were negatively impacted by COVID-19.”); 142-1 ¶ 9 (“Given the nature of the pandemic, our training opportunities for poll workers were extremely limited ahead of the June 9 primary.”).) Workers thus were not always prepared to operate the machines on Election Day. Many counties also lost “a significant number of polling places,” as well as “hundreds of regular poll workers,” due to the pandemic. (See, e.g., Dkts. 108-1 ¶¶ 5, 11; 108-2 ¶¶ 5, 7; 108-3 ¶ 7.) These disruptions were not limited to the Nine Counties or, indeed, to Georgia. “Longer than usual lines [were] reported in cities across the country,” suggesting that the pandemic was behind many of the problems and that this year’s primaries were far from normal across the board. https://www.politico.com/news/2020/06/09/georgia-primary-electionvoting-309066 (cited at Dkt. 111 at 4 n.1). The November 2020 Election 18 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 19 of 78 will also take place during the pandemic, of course. But this time Defendants have had months to prepare for it based on the experience and lessons of a prior “pandemic election.” 2. Defendants Have Made November 2020 Election Changes for the Defendants have taken steps to ensure that the long lines of June will not be repeated in November. This further reduces the predictive value of the June 2020 Primary (and earlier elections).6 a) Statewide Plaintiffs say certain counties have experienced long voting lines because they lack sufficient poll workers, technicians, backup paper pollbooks, and emergency paper ballots. Georgia officials have taken steps to address those issues across the state:  Poll Workers: The Secretary of State is partnering with the U.S. Election Assistance Commission to help counties recruit Cobb County Defendants “have made significant changes to their voting machine allocation” but do not otherwise identify specific changes for the November 2020 Election. (Dkt. 148 at 17.) Plaintiffs admit that, based on their own expert’s analysis, “Cobb County does not have any locations that are predicted to have voting lines of over 30 minutes during peak times due to voting machine allocation.” (Id.) Chatham County Defendants also have not identified specific changes for the November 2020 Election. But there is little evidence of their plans one way or the other, much less that they are insufficient. 6 19 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 20 of 78 enough poll workers. (Dkt. 110-6.) The Secretary has also established a website to facilitate recruitment of poll workers. (Dkt. 110-7.)  Technicians: Recognizing that technical issues with voting equipment contributed to problems during the June 2020 Primary, the Secretary of State has taken steps to increase technical assistance to the counties. Specifically, the state’s “vendor is increasing the number of Election Day technicians available, aiming to have a contracted technician available for each polling place.” (Dkt. 110-5 ¶ 5.) This far exceeds Plaintiffs’ request for the state to provide one technician for every ten polling places. (Dkt. 119-3 at 9.)  Backup Supplies: Voting locations will now have at least one backup paper pollbook and “a sufficient stock of emergency paper ballots” for the November 2020 Election. Curling v. Raffensperger, 2020 WL 5757809, at *25 (N.D. Ga. Sept. 28, 2020).  Additional Measures: The state has also expanded absentee and early voting options across Georgia. (See, e.g., Dkts. 110-8; 110-10; 110-15 at 6.) b) Fulton County Plaintiffs allege that long lines have repeatedly occurred at certain voting locations within Fulton County over the past decade. (Dkt. 1 ¶¶ 63–99.) Plaintiffs say these lines were caused by insufficient voting locations, insufficient voting equipment, poor planning, equipment malfunctions and insufficient technical support, understaffing of voting locations, inadequate poll worker training, problems with absentee ballot 20 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 21 of 78 voting, and insufficient emergency paper ballots. (Id.) Plaintiffs claim Fulton County Defendants will again make these mistakes in November 2020. (See, e.g., Dkt. ¶¶ 6, 98–99, 208.)7 But the evidence shows Fulton County has taken steps to address the problems Plaintiffs identify. Plaintiffs allege, for example, that Fulton County voters endured long lines in June 2020 because of consolidated polling locations, most notably the Park Tavern location where Futon County directed 16,000 registered voters. (Dkt. 1 ¶ 88; see also Dkt. 93-24 at 3 (“[M]ore than 16,000 Atlantans were assigned to vote [at Park Tavern] after two polling places backed out.”).) Fulton County has since removed two precincts (about 12,000 voters) from that location. (Dkt. 110-15 at 5.) And, beyond Park Tavern, it has reassigned more than 170,000 Fulton County voters to different polling places for the upcoming election. (Dkt. 110-5 at 1.) Indeed, the county has dramatically increased its voting locations from 164 in June 2020 to 255 in November 2020. (Dkt. 147-1 ¶ 5.) Only 4 of these locations have more than 5,000 assigned voters, meaning the county now “has fewer voters Plaintiffs generally allege that each Defendant will repeat their past mistakes in the upcoming November 2020 Election. (See, e.g., Dkt. 1 ¶¶ 6, 208.) 7 21 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 22 of 78 assigned per precinct than at any time during the tenure of its current Elections Department director.” (Id. ¶ 6.) A senior Fulton County election official provided other evidence the county has “made significant changes to address [past] issues and to facilitate a better voting experience in the November election.” (Dkt. 1471 ¶ 4.) Those changes include the following:  Voting Equipment: Fulton County has “purchased millions of dollars of voting equipment” and now has at least one BMD for every 250 registered voters. (Id. ¶ 8.) It has also purchased two mobile voting units to help alleviate lines at any voting locations requiring assistance. (Id. ¶ 9.) Plaintiffs concede that Fulton County has “made significant changes to their voting machine allocation” and that this puts the county at “lower risk” of long lines in November. (Dkt. 148 at 17.)  Technical Issues: Fulton County has assigned a technician to each polling location (to arrive by 5 a.m.), ensured each location can accommodate its voting equipment, deployed a new electronic communication system to ensure voting locations can obtain “additional technological help,” and has purchased items to reset poll pads if they experience problems. (Dkt. 147-1 ¶ 10.)  Poll Workers: Fulton County has “revamped” its poll worker training to “provide more hands-on opportunities to prepare equipment and to understand the procedures.” (Dkt. 108-2 ¶ 8.) It has also hired 510 line managers and created an additional call center to handle election day calls from the voting locations. (Dkt. 147-1 ¶ 10.)  Voting Alternatives: Fulton County has expanded early voting and absentee voting. It will have 30 polling locations — including 3 “mega sites” — open for early voting for 3 weeks 22 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 23 of 78 (including weekends). (Id. ¶ 7.) It has created an online portal for absentee ballot applications, almost doubled its number of absentee ballot drop boxes, and created a smartphone application to help voters with several aspects of voting. (Id. ¶¶ 11–13.)  Emergency Paper Ballots: Fulton County has doubled the amount of emergency paper ballots at each voting location, exceeding the state law requirement by a factor of two. (Id. ¶ 10). c) DeKalb County Plaintiffs allege that long lines occurred at certain voting locations within DeKalb County in 2008, 2016, 2018, and June 2020. (Dkt. 1 ¶¶ 100–115.) Plaintiffs say these lines were caused by the consolidation and relocation of voting locations, equipment malfunctions, insufficient technical support, understaffing, insufficient training of poll workers, and insufficient emergency paper ballots. (Id.) But a senior DeKalb County election official has presented evidence that the county is taking steps to shorten lines for the November 2020 Election “based on lessons learned and experience from the June 2020 primary and August 2020 run-off election.” (Dkt. 143-1 ¶ 6.) Those steps include the following:  Voting Locations: Only 2 of DeKalb County’s 176 precincts now have more than 5,550 registered voters. (See Dkt. 127-1.) Where the pandemic has required precincts to be consolidated into a single voting location, the county will maintain “separate 23 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 24 of 78 voting lines, registration, machines and poll workers for each precinct.” (Dkt. 143-1 ¶ 4.) The county does not expect “any shortage of polling places.” (Id. ¶ 5.)  Technical Issues: In addition to the technicians provided by the state, the county will provide at least 20 technicians of its own. (Id. ¶ 15.)  Poll Workers: DeKalb County has “increased recruitment efforts for poll workers” and has now secured 1,800 workers as well as “additional back-up poll workers” in case of cancellations. (Id. ¶¶ 9–10.) The county “expects there to be a sufficient amount of poll workers staffed at each of the polling places.” (Id. ¶ 8.)  Training: DeKalb County “has already commenced training poll workers” (including backup workers) and “is providing improved training” for the November 2020 Election. (Id. ¶¶ 10– 11.) This will include “in-person training” and “a live simulation of the entire voting process from start to finish.” (Id. ¶ 11.)  Emergency Paper Ballots: DeKalb County expects to have “a sufficient amount of emergency paper ballots” at each voting location, and “has the capacity to print [and deliver] additional emergency paper ballots on Election Day” if necessary. (Id. ¶¶ 8, 13.) The ability to print emergency paper ballots goes directly to Plaintiffs’ claim that County Defendants have insufficient paper ballots at the polling locations.  Voting Alternatives: DeKalb County will have 12 early voting locations open for almost 3 weeks (including weekends). (Id. ¶ 3.) The county also has “taken steps to improve the efficiency of the absentee voting process,” including by acquiring new scanning equipment and engaging a non-profit organization with expertise in absentee voting. (Id. ¶ 14.) The county is sending absentee ballot applications to all registered voters and is placing 24 absentee ballot boxes throughout the county. (Id. ¶¶ 17–18.) 24 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 25 of 78  Line-Tracking: DeKalb County will deploy a new software application that allows voters and the county to monitor line length at all voting locations. (Id. ¶ 16.) d) Gwinnett County Plaintiffs allege long lines occurred at certain voting locations within Gwinnett County in 2016, 2018, and June 2020. (Dkt. 1 ¶¶ 129– 142.) Plaintiffs say these lines were caused by polling place consolidation, technical issues, understaffing, inadequate training of poll workers, and the delayed delivery of equipment to voting locations. (Id.) But a senior Gwinnett County official presented evidence that the county “has taken a number of steps to learn from the June primary and prepare for the November election.” (Dkt. 108-1 ¶ 5; see Dkt. 142-1 ¶ 12.) Those steps include:  Voting Equipment: Gwinnett County has “made significant changes to their voting machine allocation,” which Plaintiffs concede puts them at “lower risk” of long lines. (Dkt. 148 at 17.)  Poll Workers: Gwinnett County has “added more in-person training opportunities for poll workers who will be working in the November election.” (Dkt. 142-1 ¶ 9; see Dkt. 108-1 ¶¶ 12– 16.) That training is now well underway. (Dkt. 108-1 ¶¶ 15–16.) The county also has “added additional individuals for November to ensure that all polls will be adequately staffed.” (Dkt. 142-1 ¶ 10.) 25 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 26 of 78  Equipment Delivery: Gwinnett County has secured “more trucks and drivers” for equipment delivery and established a “better communication” system with its transportation contractor. (Id. ¶ 5.) The county also has revised its “instructions for [equipment] logic and accuracy testing” in advance of delivery. (Id. ¶¶ 7–8.) The county implemented these changes in an August 2020 election and “all equipment was delivered and set up well in advance of the start of voting.” (Id. ¶¶ 5, 8.) The county expects no delivery issues in November 2020. (Id. ¶ 6.)  Alternative Voting: “For the first time ever, Gwinnett County will open nine early-voting locations for all three weeks of early voting.” (Id. ¶ 11.) e) Clayton County Plaintiffs allege long lines occurred at certain voting locations within Clayton County in 2008, 2018, and June 2020. (Dkt. 1 ¶¶ 157– 167.) Plaintiffs say these lines were caused by the consolidation of voting locations, insufficient poll workers, inadequate training, technical issues, and insufficient emergency paper ballots. (Id.) But a senior Clayton County official has provided an affidavit showing the county is making changes for the November 2020 Election. (Dkt. 140.) Those changes include:  Voting Locations: Clayton County has increased its voting locations from 58 in June 2020 to 65 in November 2020. (Id. ¶ 4.) The county is monitoring early and absentee voting so it can adjust its equipment allocations based on unanticipated voting patterns. (Id. ¶ 10.) 26 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 27 of 78  Poll Workers: Clayton County is providing in-person training to poll workers. (Id. ¶ 12.) Training is offered 6 days per week, several times per day. (Id.) The county has reduced class sizes to 20 workers. (Id.) f) Henry County Plaintiffs allege long lines occurred at certain voting locations within Henry County in June 2020. (Dkt. 1 ¶¶ 168–174.) Plaintiffs say these lines were caused by the consolidation of voting locations, understaffing, inadequate training, and technical issues. (Id.) But a senior election official provided evidence the county is making changes “based on lessons learned and experience from the June 2020 primary and August 2020 run-off election.” (Dkt. 108-3 ¶ 6.) Those changes include:  Voting Equipment: Henry County has secured 55% more poll pads, 117% more ballot marking devices, and about 30% more scanners than it had for the June 2020 Primary. (See Dkt. 1491 at 9.)  Poll Workers: Henry County “has increased recruitment efforts for poll workers” by working with the state and other local organizations. (Dkt. 108-3 ¶ 7.) It has now secured “a sufficient number of poll workers” for each polling place, as well as backup workers in case of cancellations. (Id.)  Training: Henry County is “improving training” for poll workers by requiring each worker to “perform[] a live simulation of the entire voting process” before the election. (Id. ¶ 9.) The 27 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 28 of 78 county will provide this training to backup workers as well. (Id.) The training program has already begun and will last for more than 3 weeks. (Id. ¶ 8.)  Backup Paper Supplies: Henry County expects to have “a sufficient amount of emergency paper ballots,” and will keep a paper list of voters at each voting location in case electronic poll pads malfunction. (Id. ¶¶ 6, 10.) g) Douglas County Plaintiffs allege long lines occurred at three voting locations within Douglas County in June 2020 due to technical issues. (Dkt. 1 ¶¶ 175– 181.) Douglas County has since determined that those issues were largely caused by insufficient “on-site technical support” and human error by an election manager. (Dkt. 146-1 ¶¶ 5–6.) The county has replaced the manager responsible and secured “an on-site technician for every polling location during the November 2020 election.” (Id.) The county also has taken other steps to ensure a smooth election in November:  Voting Equipment: It has “increased the number of poll pads, BMDs, and ballot scanners to many of its precincts.” (Dkt. 1461 ¶ 7.)  Alternative Voting: It has increased its early voting locations from 5 in June 2020 to 8 in November 2020 — and it has notified voters of these new locations by email. (Id. ¶ 4.) The county also has mailed absentee ballot applications to all registered voters. (Id.) 28 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 29 of 78 h) Macon-Bibb County Plaintiffs allege long lines occurred at certain voting locations within Macon-Bibb County in 2012, 2016, 2018, and June 2020. (Dkt. 1 ¶¶ 182–189.) Plaintiffs say these lines were caused by insufficient training, technical issues, understaffing, and delays setting up equipment on Election Day. (Id. ¶¶ 183–186.) But Macon-Bibb County has taken steps to address these issues and to otherwise ensure a smooth election in November:  Set-Up Delays: Poll workers will now arrive earlier to set up the voting equipment. (Dkt. 128-1 ¶ 5.)  Staffing: Macon-Bibb County has secured 357 poll workers for its 31 voting locations. (Dkt. 128-3.) It says it “will have enough . . . poll workers to accommodate each and every polling location in” the county. (Dkt. 128-1 ¶¶ 9–10.)  Training: Macon-Bibb County is now conducting “in-person, hands-on poll worker training.” (Id. ¶ 6.) Training for the June 2020 Primary was “almost exclusively virtual.” (Id. ¶ 6.) i. Conclusion The evidence shows Defendants have taken extensive measures to address the issues that caused long lines in the past. It is possible, of course, these measures will ultimately prove insufficient and long lines will still arise. But that is not the point; no one, including this Court, can 29 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 30 of 78 guarantee short lines. See Georgia Shift v. Gwinnett Cty., 2020 WL 864938, at *5 (N.D. Ga. Feb. 12, 2020) (“This Court cannot guarantee that voters will not have to stand in line.”). The point is these measures materially distinguish the November 2020 Election from the June 2020 Primary (and earlier elections) — and this precludes the Court from assuming June (or the more distant past) will repeat itself in November. 3. Plaintiffs’ Cases are Unavailing Plaintiffs cite O’Shea, 414 U.S. 488 and Lynch v. Baxley, 744 F.2d 1452, 1456 (11th Cir. 1984) in support of their theory that Defendants’ past conduct establishes imminent injury. (Dkt. 111 at 4–5.) But the O’Shea court found there was not standing, rejecting the argument that defendants’ “pattern and practice of conduct” established imminent injury. O’Shea, 414 U.S. at 493–99. And although the court acknowledged that “past wrongs are evidence bearing on whether there is a real and immediate threat of repeated injury,” it also said that “[p]ast exposure to illegal conduct does not in itself show a present case or controversy regarding injunctive relief.” Id. at 495–96. Lynch is equally unhelpful. The court there held plaintiff was “realistically threatened by a repetition of his experiences and therefore 30 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 31 of 78 ha[d] standing to seek the requested injunction.” 744 F.2d at 1457 (emphasis added). But Lynch was decided years before the Supreme Court expressly adopted the heightened standing requirement of a “certainly impending” injury. Whitmore v. Arkansas, 495 U.S. 149, 158 (1990) (“A threatened injury must be certainly impending to constitute injury in fact.”). And, more recently, the Court has squarely rejected an attempt to “replace the requirement of imminent harm . . . with the requirement of a realistic threat” of injury — the very standard on which the Lynch court relied. Summers, 555 U.S. at 499–500; see Georgia Shift, 2020 WL 864938, at *3 (“In Summers, the Supreme Court rejected a standing test that would replace the requirement of ‘imminent’ harm with the requirement of ‘a realistic threat.’”). Given these developments, Lynch is not useful guide here.8 Lynch is also distinguishable on the facts. The plaintiff there sought to enjoin the state from “detaining in county jails persons awaiting mental illness involuntary commitment proceedings.” 744 F.2d at 1454. The court said he had standing including because (1) he had been incarcerated twice in the last three years while he awaited commitment proceedings, and (2) the state had “ordered the possibly mentally ill detained in three-fourths of the county jails in the state.” Id. at 1456–57. But the court did not focus only on historical facts. The court emphasized that plaintiff suffered from an existing “mental condition,” that a statute expressly authorized the state to incarcerate those awaiting commitment 8 31 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 32 of 78 Plaintiffs’ strongest authorities are actually cases they fail to cite: Ciudadanos, 622 F.2d 807 and 31 Foster Children v. Bush, 329 F.3d 1255 (11th Cir. 2003). But even those cases do not establish imminence here. Ciudadanos involved a Texas statute that left the “selection of potential grand jurors . . . entirely to the discretion of the jury commissioners.” 622 F.2d at 811. Plaintiffs sued the commissioners, claiming they systematically excluded members of plaintiffs’ demographic group from jury selection in violation of the Fourteenth Amendment. Id. at 812–13. The court found plaintiffs had standing to seek injunctive relief because (1) “the very nature of the Texas selection scheme” was “highly subjective and susceptible to abuse”; (2) the commissioners had, for “over ten years[,] . . . consistently produced grand juror lists upon which the classes to which [plaintiffs] belong have been substantially underrepresented”; proceedings, that the state admittedly continued to invoke the statute, that “mental health facilities are not available in certain counties,” and that jails were the only facilities “routinely available” for many people awaiting commitment proceedings. Id. All of these facts were present and ongoing, rather than historical, making them plausibly predictive of future harm. In contrast, Plaintiffs here focus largely on historical facts — long lines at past elections. They say little about present and ongoing facts that make long lines likely in the future. 32 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 33 of 78 and (3) this evidence was “sufficient to give rise to a strong inference that the injury will be repeated in the future.” Id. at 820–21. 31 Foster Children involved a putative class action brought by several children in Florida’s foster care system. 329 F.3d at 1260. The children claimed that Florida’s foster care officials consistently provided them with terrible care in violation of the U.S. Constitution. Id. at 1260– 61, 1264–65. The court found the children established an imminent injury because (1) “[t]hey cannot avoid exposure to the defendants’ challenged conduct” and (2) “[t]he alleged systemic deficiencies in the Florida foster care system are similar to an injurious policy.” Id. at 1266. Ciudadanos and 31 Foster Children stand for the proposition that past wrongs can predict future wrongs if there is a long record of wrongs and no reason to think they will not continue. But that is not our case. Defendants here have taken substantial steps to address the issues that caused long lines in the past. So the Court cannot assume those issues will continue in the future based only on their existence in the past — especially when there are unique 33 elements about the past Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 34 of 78 (e.g., the intersection of a new voting system and a new pandemic).9 The bottom line is that past is not necessarily prologue here. The Court rejects Plaintiffs’ argument that, because long lines have occurred in the past, they are certainly impending in November 2020.10 D. Plaintiffs’ Expert Predictions Do Not Establish a Certainly Impending Injury To bolster their contention that Defendants are likely to repeat the failures of the past, Plaintiffs have retained Dr. Muer Yang to predict November 2020 Election Day wait times at polls in seven of the Nine Counties.11 Dr. Yang is an associate professor of operations management at a university in Minnesota. (Dkt. 93-62 at 2.) His predictions are based largely on the number of voters assigned to each polling location and the amount of voting equipment available to accommodate those voters. As explained above, Georgia’s pre-2020 elections are also distinguishable from today’s elections because they used a different voting system. 10 Church v. City of Huntsville, 30 F.3d 1332 (11th Cir. 1994) also relied on past wrongs to predict future wrongs. But it has little in common with our case. In Church, unlike here, there was no reason to believe defendants would change their conduct going forward. Indeed, there was every reason to believe they would continue their conduct because it was part of “a well organized and coordinated campaign” — an intentional and ongoing municipal policy — to cause the injuries about which plaintiffs complained. Id. at 1339. 11 Dr. Yang offers no prediction for Chatham County or Clayton County. 9 34 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 35 of 78 Dr. Yang initially submitted a declaration in which he offered predictions for only Henry and Fulton Counties (“First Report”). (Dkt. 93-62 at 40–45, 48–51.) But his analysis was based on the counties’ equipment allocation during the June 2020 Primary and, as explained above, both counties are making significant adjustments to alleviate the problems they experienced in that election. As mentioned above, in an effort to afford Plaintiffs the opportunity to show the requisite certainty of a future injury for standing, the Court (at the close of its first hearing) ordered County Defendants to provide Dr. Yang with additional data so he could perform his analysis for all Nine Counties based on their equipment and voter allocations for the November 2020 Election. Specifically, the Court ordered County Defendants to provide Plaintiffs the number of registered voters assigned to each polling location and the number of BMDs, scanners, and poll pads assigned to each of those locations for both the June 2020 Primary and the upcoming November 2020 Election. (See Dkts. 125; 138; 159 at 119–127.) This was the very information Plaintiff’s sought in their preliminary injunction motion so that Dr. Yang could calculate wait times at each polling place in November. (Dkt. 119-3 at 5.) 35 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 36 of 78 Dr. Yang has now submitted a second declaration, redoing his analysis for Henry and Fulton Counties and offering new predictions for the remaining counties (“Second Report”). (Dkt. 105-1.) Dr. Yang predicts wait times of at least 30 minutes — which he defines as “long” — at certain voting locations within Fulton County, Gwinnett County, Macon-Bibb County, Henry County, Douglas County, and DeKalb County. Plaintiffs say this shows long lines are certainly impending at those locations. The Court disagrees.12 In their complaint and personal declarations, Plaintiffs repeatedly say their injuries will be triggered by “long lines.” (See Dkts. 1 ¶¶ 12–16; 9330 ¶ 9; 93-31 ¶ 8; 93-32 ¶ 8; 93-60 ¶¶ 10, 18–19; 93-63 ¶¶ 10–11.) They never say (at least for the purposes of their alleged injury) that “long” means “30 minutes”; indeed, they often suggest it means something closer to hours. (See, e.g., Dkts. 1 ¶¶ 12–14 (complaining about hourslong lines); 93-30 ¶ 9 (“I am concerned I will have to wait in hours long lines again.”); 93-31 ¶ 8 (“I am concerned about . . . wait[ing] in line for hours to cast my vote.”); 93-32 ¶ 8 (“I do not know if I will be able to stand in line for eight hours” and “[s]tanding in line . . . for hours puts me at risk of contracting COVID-19”); 93-60 ¶¶ 6–9 (referring to “[l]ong lines with up to two-hour wait times,” three-hour waits, “voters waiting for several hours,” and “eight-hour lines”); 93-63 ¶ 9 (referring to “voters wait[ing] up to eight hours”).) There is no evidence or allegation that Individual Plaintiffs are worried about standing in a line that is 31 minutes instead of 29 minutes. Nor is there any evidence or allegation that Organizational Plaintiffs will decide whether to divert resources based on that distinction. It is unclear whether Plaintiffs may now, postpleading, redefine their injury in the way they seek to do. But, for the purposes of assessing Plaintiffs’ evidence, the Court will consider Dr. Yang’s 30-minute barrier. 12 36 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 37 of 78 1. Dr. Yang’s Methodology As explained earlier, Georgia uses a three-stage voting process: (1) voters check in using electronic poll pads; (2) they cast their ballots on BMDs; and (3) they deposit their printed ballots through scanners. The following graphic illustrates how the process works: (Dkt. 93-62 at 11.) In his analysis, Dr. Yang predicts voters’ wait times at each stage of the voting process (check in, cast ballot, deposit ballot) and then adds those waits together to predict voters’ total wait time. (Dkt. 93-62 at 11.) 37 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 38 of 78 He does this for each voting location within the counties he studied. He estimates both the average wait time and the maximum wait time at each location.13 He calculates two average wait times for each location: one, using a queue formula known as M/M/c; and the other, using a simulation model developed by someone called Mark Pelczarski. He also uses the simulation model to calculate the maximum wait time at each location. Dr. Yang’s predictions (whether calculated under M/M/c or the simulation model) are based on four main inputs: (1) the number of voters expected to vote at each voting location on Election Day in November 2020, (2) the rate at which voters will arrive at those locations throughout the day, (3) the amount of voting equipment allocated to each location, and (4) how long it will take voters to use that equipment once they access it. (See Dkt. 93-62 at 11–12.) Notice that inputs (1), (2) and (4) depend on “how independent decisionmakers will exercise their judgment” in the November 2020 “Average wait time at a polling place is the sum of every voter’s wait time divided by the total number of voters turned out at this poll.” (Dkt. 93-62 at 6 n.5.) “The maximum wait time is the average of wait times of voters arriving during the worst 15-minute intervals throughout the day.” (Dkt. 149-1 at 11 n.16.) 13 38 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 39 of 78 Election: whether thousands of individual registered voters will decide to vote; if so, using what method; if in person, on what day; if on Election Day, at what time and how long on each machine. Clapper, 568 U.S. at 413. As a matter of law (rather than science), this raises immediate red flags because standing is “substantially more difficult to establish” where it “depends on the unfettered choices made by independent actors not before the courts and whose exercise of broad and legitimate discretion the courts cannot presume either to control or to predict.” Lujan v. Defs. of Wildlife, 504 U.S. 555, 562 (1992); see Clapper, 568 U.S. at 413 (“[W]e have been reluctant to endorse standing theories that require guesswork as to how independent decisionmakers will exercise their judgment.”). Moreover, any “theory of standing, which relies on a highly attenuated chain of possibilities, does not satisfy the requirement that threatened injury must be certainly impending.” Clapper, 568 U.S. at 410. There is no doubt that Dr. Yang’s predictions do rely on a chain of possibilities or a sequence of uncertain assumptions. That, too, raises red flags. 39 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 40 of 78 2. General Problems with Dr. Yang’s Predictions In addition to the legal red flags mentioned above, the Court has several concerns with Dr. Yang’s methodology and assumptions. First, he never fully explains the M/M/c queue theory or the simulation model on which his predictions are based. He says, in a footnote in his reports, that the M/M/c model “assumes that a voter’s arrival process follows a Poisson distribution, the service time follows an exponential distribution, and c servers in the queue system.” (Dkts. 93-62 at 9 n.7; 149-1 at 4 n.6.) That is not exactly illuminating. He says “[s]imulation models are another popular method used to study waiting lines” because they “capture more dynamic characteristics of queues” (some of which he then identifies). (Dkt. 149-1 at 2.) But that hardly tells the Court what the simulation model actually is. There are, of course, other insights into both models scattered throughout his reports — for example, that M/M/c assumes “voters will arrive at a constant rate throughout the day” whereas the simulation model does not — but no ordinary person can walk away from Dr. Yang’s report feeling clear about either approach. (Id.) That suggests the Court should be cautious before deferring to them. 40 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 41 of 78 Second, Dr. Yang assumes voters will spend 45 seconds checking in with a poll pad and 20 seconds depositing their printed ballot in the scanner. (Dkt. 149-1 at 7.) But he assumed a check-in time of 30 seconds in his First Report and does not explain why he now believes it will be 50% longer. (Dkt. 93-62 at 12.) He admits “a small number of polling places will be affected by this 15-second increase” — meaning it causes them to exceed his 30-minute barrier — though he does identify which places those are. (Dkt. 170-2 at 26.) Dr. Yang also assumed, in his First Report, that voters would spend 14 seconds depositing their ballots in the scanner. (Dkt. 93-62 at 49.) That assumption was based on a video demonstration of the voting process, which showed a voter spending 14 seconds at the scanner. (Dkt. 93-62 at 12.) He does not explain why he now believes that time will be 43% longer. Nor does he say how much, or at what locations, this increase affected his projections.14 Dr. Yang also uses a formula, based on voting data in South Carolina, to predict the amount of time voters will spend using a BDM to cast their ballot. (Dkt. 170-2 at 26–27.) Although he says the “overall logic behind th[e] model” applies to Georgia, he says “local election officials” might determine the real number is “shorter” or “longer” based on circumstances in their own jurisdictions. (Id. at 27–28.) 14 41 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 42 of 78 Third, there are serious problems with Dr. Yang’s calculation of expected in-person voters on Election Day in November 2020 — a key input on which his calculations are based. In his First Report, he assumed for Henry County that roughly the same percentage of registered voters who voted on Election Day in November 2016 would do so again in November 2020 (minus an anticipated increase in absentee voting due to COVID-19). (Dkt. 93-62 at 40–41.) Now he says the number will be 40% of the total voter turnout in November 2020 (“40% Methodology”). (Dkt. 149-1 at 8.)15 He gets this 40% figure from an online national poll (dated July 2020) that asked: “Thinking ahead to the (2020 presidential) election in November, would your preference be to vote in person on Election Day, vote in-person early, or vote by mail?”16 (Dkt. 149-1 at 8.) But somebody’s preference for a particular voting method says little about the method they intend to use. No doubt lots of people would prefer to vote in person. But many will not do so because His Second Report applies the 40% Methodology to all counties other than Douglas County. For that county, he uses projections apparently provided by the Secretary of State. There is no evidence of how those projections were calculated. (Dkt. 170-2 at 29–30, 35–36, 86–87.) 16 The poll is available at https://ropercenter.cornell.edu/psearch/question_view.cfm?qid=1958132 &pid=50&ccid=50#top. 15 42 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 43 of 78 other considerations — like the pandemic or personal responsibilities — outweigh that preference. (See Dkt. 93-62 at 40 (“Due to the Covid-19 pandemic, voting by mail is expected to be much higher than in previous elections.”); Dkt. 170-2 at 95.) Dr. Yang then applies his 40% figure to the total number of people expected to vote in November 2020. To determine this latter number, he applies what he believes to be the overall turnout rate at each precinct in November 2016 to the total number of registered voters assigned to that precinct today. He gets the November 2016 turnout rate from the Georgia Secretary of State’s website. (Dkt. 149-1 at 8 n.13.) But that website appears to report conflicting rates for each county. And, although Dr. Yang does not know which rate is accurate, his report generally uses the higher one. (See Dkt. 170-2 at 40–44, 48–50.)17 One set of rates is located at https://sos.ga.gov/index.php/Elections/current_and_past_elections_resul ts; another at https://sos.ga.gov/index.php/elections/general_election_turnout_by_dem ographics_november_2016. For example, the former says turnout rate was 75% in Fulton County in November 2016. But the latter says that rate was 57%. Dr. Yang includes the higher rate (75%) in his report. (Dkt. 149-1 at 22.) There are similarly conflicting numbers for Douglas County (78% vs. 63%), Macon-Bibb County (72% vs. 60%), and other 17 43 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 44 of 78 Dr. Yang testified at his evidentiary hearing that he now prefers his second methodology for predicting election day voters in November 2020 (40% x projected overall turnout) because the COVID-19 pandemic makes the November 2016 data less reliable. (See, e.g., Dkt. 170-2 at 44– 45, 84–85.) But this does not account for the unreliability of the 40% number itself (pulled from a poll about preferences rather than plans or past performance). Nor is the assertion credible since (1) Dr. Yang’s First Report used the very methodology he now criticizes and (2) his Second Report still relies on data (overall turnout) from the November 2016 Election. Indeed, even at his evidentiary hearing, Dr. Yang testified that “you can pick [his first methodology] to project election voters this year.” (Dkt. 170-2 at 50; see also id. at 61 (“[Y]ou can argue either way.”), 72 (“[Y]ou can probably do that.”).) He further testified that the 40% Methodology “was just presenting you [with] another case” but that it was up to officials’ “local elections judgment” to decide how “likely” that scenario was. (Dkt. 170-2 at 72.) The problems with Dr. Yang’s 40% counties. (See Dkt. 149-1 at 17, 25 (Dr. Yang citing the higher rate for Douglas and Macon-Bibb Counties).) 44 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 45 of 78 Methodology are explored in more detail in the Court’s analysis of Henry County below. 3. Dr. Yang’s Predictions Fall Short in Each County The Court now looks at Dr. Yang’s specific predictions for the counties he studied. His predictions ultimately establish no more than a possibility — not even a reasonable likelihood — of long voting lines in November 2020. Under binding precedent, that is not enough to show standing. a) Cobb, Fulton, Gwinnett, and Macon-Bibb Counties Having received Dr. Yang’s Second Report, Plaintiffs concede that “Cobb County, Fulton County, and Gwinnett County have made significant changes to their voting machine allocation and, due to those re-allocations, appear to be at a lower risk for voting lines of over 30 minutes long, at least due to machine allocation.” (Dkt. 148 at 17.) Although Plaintiffs do not say so explicitly, Macon-Bibb County is clearly “lower risk” as well since its projected wait times are even lower than Fulton County’s (and about the same as Gwinnett’s). (Dkt. 149-1 at 26.) Plaintiffs admit Cobb County “does not have any locations that are predicted to have voting lines of over 30 minutes during peak times due 45 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 46 of 78 to voting machine allocation.” (Dkt. 148 at 17.) So Dr. Yang’s analysis offers no support for Plaintiffs’ contention that Cobb County voters face a “certainly impending” injury from long lines. Dr. Yang also found that Fulton County allocated enough BMDs to each voting location. (Dkt. 149-1 at 23.) But he did raise concerns about the number of scanners at two Fulton County locations: 1272 and 729 (also identified as 08E/09F and ML03/ML07AB). (Id. at 23.) He concluded that the county’s decision to have just one scanner at each of these locations could lead to wait times of 39 and 31 minutes during peak periods. (Id.) He thus recommended the addition of another scanner at each location. (Id.) During the evidentiary hearing, Fulton County Defendants pointed to their most recent equipment allocation plan, which shows three scanners at locations 1272 and 729. (Dkts. 168-1 at 1; 170-2 at 75–78.) Dr. Yang conceded he had been unaware of this allocation, that it exceeded his recommendation, and that he no longer has any basis to believe the county’s allocation of voting equipment will result in wait times (even during peak periods) that exceed his 30-minute barrier. (Id.) 46 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 47 of 78 Finally, Dr. Yang found that Gwinnett and Macon-Bibb Counties each have two voting locations at which there may be maximum wait times of 30–32 minutes. (Dkt. 149-1 at 24–26.)18 But he testified at his evidentiary hearing that there was “wiggle room” in his numbers — which were not “precise” — and that a projected wait time of 31 minutes “very possibl[y] . . . could be 29 minutes or about 28 minutes” or even 25 minutes. (Dkt. 170-2 at 23, 79.) So his predictions for Gwinnett and Macon-Bibb Counties — especially when combined with the Court’s general concerns outlined above — establish no more than “a ‘perhaps’ or ‘maybe’ chance” of 30-minute lines in those counties. County Bowen, 233 F.3d at 1340. That is insufficient to show standing, even assuming 30 minutes constitutes a “long line.” Id. b) Henry County Henry County Defendants have perhaps the most compelling argument against standing of all the counties for which Dr. Yang offered predictions. In his First Report, Dr. Yang analyzed Henry County’s voter and equipment allocation information for the June 2020 Primary. He He conceded the average wait times at these locations would be no more than 9 minutes, raising a concern only about the busiest 15 minutes in the voting day. (Dkt. 149-1 at 24, 26.) 18 47 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 48 of 78 found that the allocation was insufficient at certain voting locations. He then recommended a new allocation plan for November 2020. (Dkt. 93-62 at 40–45.) He said his plan would result in average wait times of “no more than three minutes long at any polling location[] and no voter would have to wait for more than 30 minutes.” (Id. at 43.) During litigation before this Court, Henry County presented evidence that, based on its experience in the June 2020 Primary, it increased and reallocated voting equipment at each of its 37 polling places for the November Election. (Dkt. 164-1 ¶ 7.) As part of this, it more than doubled the number of BMDs it plans to deploy from 222 in June to 482 in November. (Id.) It also increased the number of poll pads it will deploy to the precincts from 74 to 115. (Id.) Henry County’s reallocations and deployment of additional voting equipment exceeded Dr. Yang’s recommendations in every precinct. (Compare Dkt. 93-62 at 44–45 with Dkt. 133-1.) For example, while Dr. Yang recommended that Henry County allocate 8 BMDs to the Locust Grove location, the county decided to include 12 machines at that location. (Dkts. 93-62 at 44; 1331.) It will send 16 BMDs to the Westside location, rather than the 8 he recommended. (Id.) The Lowes precinct will get 20 rather than the 12 48 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 49 of 78 he recommended, and the North Hampton and Stockbridge precincts will get 20 and 16 rather than the 9 and 8 he proposed. (Id.) After receiving Henry County’s allocation plan for the November 2020 election, Plaintiffs asked Dr. Yang to reanalyze the county’s deployment of voting equipment. As the county has now exceeded Dr. Yang’s prior recommendations, one might expect his new analysis to re-verify his prior conclusion that “no voter would have to wait for more than 30 minutes.” But, not so. In his Second Report, Dr. Yang dramatically changed his methodology to increase his estimate of Election Day turnout — and that change allowed him to predict maximum wait times of 31–91 minutes in 7 Henry County precincts, notwithstanding his earlier prediction that none of those precincts would have problems in November. (Dkt. 149-1 at 11.) In his first analysis, he relied on November 2016 voter data from the Georgia Secretary of State’s website to conclude that “[a]bout 22% of the total registered voters are Election Day voters” in Henry County. (Dkt. 93-62 at 40.) In recognition of his belief that “Covid-19 will increase the mail-in absentees in the 2020 General Election” like it did in the June 2020 Primary, he ultimately assumed that “Election Day voters at each 49 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 50 of 78 polling place in Henry County in the November General Election [would be] 19% of the total registered voters.” (Id. at 40–41.) In his second analysis, he threw out his 19% in-person Election Day rate. Instead, he determined from documents on the Georgia Secretary of State’s website that Henry County had a total voter turnout rate of 79% in the November 2016 Election. (Dkts. 149-1 at 8 & n.13; 170-2 at 59–60.) He then considered a July 2020 poll from the Pew Research Center’s American Trends Panel which found that, when “[t]hinking ahead to the (2020 presidential) election in November,” 40% of respondents said their “preference” would be to vote in person on Election Day. (Dkt. 149-1 at 8.) He accepted this as showing that “40% of voters are likely to vote in person on Election Day.” (Id.) Applying these two multiples together (79% x 40%), he concluded that 31.6% of Henry County’s registered voters would vote on Election Day next month — a 12.6% increase from his prior estimate. (Dkt. 170-2 at 59–60.) He relied on this calculation despite no evidence suggesting Henry County voters had ever before turned out to vote on Election Day at that level. He did this despite his previous calculation using actual historical data showing only 22% of the county’s registered voters go to the polls on 50 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 51 of 78 Election Day. He also did this despite having “assume[d] for the purposes of [his second] analysis that the overall voter turnout in the 2020 Presidential Election [will be] the same as the turnout for the 2016 Presidential Election, though it may in fact be higher.” (Dkt. 149-1 at 8.) Dr. Yang’s decision to increase his assumption of Election Day voting by 12.6% had a profound impact on his conclusions. For the Locust Grove precinct, for example, his Second Report estimated 2,275 Election Day voters, nearly doubling his previous estimate of 1,237 voters. (Dkts. 93-62 at 44; 149-1 at 11.) His estimate for voter turnout at the Sandy Ridge precinct also nearly doubled, increasing from his initial estimate of 896 to 1,641 (an increase of 745 voters). (Id.) He had similarly dramatic increases in several other Henry County precincts. (Id.) As a result of his methodological change, Dr. Yang now predicts maximum wait times exceeding his 30-minute barrier at 7 precincts. He recommends adding more equipment to those locations. Dr. Yang’s new analysis is too speculative to show that 30-minute lines — much less “long lines” — are certainly impending in Henry County next month. His analysis turns on the 40% figure taken from the Pew poll and the November 2016 overall turnout rates reported on the 51 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 52 of 78 Secretary’s website. The Court finds both suspect. As to the first, Dr. Yang said he thought the Pew poll presents the “best data” he could find. (Dkt. 170-2 at 37.) But he could provide no credible reason for abandoning his prior determination that 19% of registered voters will vote on Election Day, particularly when he continues to assume throughout his report that total turnout in the November 2020 Election will be the same as the November 2016 Election. He testified that the COVID-19 outbreak “might” have some “kind of influence on the voter turnout this year.” (Dkt. 170-2 at 45; see id. (“[I]f there was no COVID19 this year, I think I would just directly go to 2016 and to [sic] use that number.”).) He suggested it could increase or decrease voter turnout, he was not sure. (Id.) He looked to the Pew poll because it was more current and, he assumed, included any impact the outbreak might have on voter preference. (Id.) It is hard for the Court to believe the COVID-19 outbreak, and the social distancing it mandates, will lead to higher in-person voter turnout, let alone that it will do so to the extent he now predicts. Indeed, in his First Report, Dr. Yang was clear that “Covid-19 will increase mail-in absentee ballots in the 2020 General election” just as it did in the June 2020 Primary, and that “voting by mail is expected 52 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 53 of 78 to be much higher than in previous elections.” (Dkt. 93-62 at 40.) He never suggested the outbreak would increase in-person voting. And County Defendants have presented plenty of evidence that they anticipate more mail-in voting as a result of the pandemic. (See, e.g., Dkts. 128-1 ¶ 9; 164-1 ¶ 6; 167-1 ¶ 7.) The Pew poll also was not conducted in Georgia, let alone Henry County. And, on its face, the poll merely asked (in July) how people would “prefer” to vote in November. It does not even purport to estimate how people intend to vote, the purpose for which Dr. Yang adopted it. At a hearing, Dr. Yang also could not confirm the accuracy of the November 2016 turnout rates on which he relied to predict total voter turnout in November 2020. While talking about his analysis, Dr Yang explained that he obtained voter turnout rate information for November 2016 from the Secretary of State’s website. (Dkt. 170-2 at 39–42.) He provided a link to the data in his report. (Dkt. 149-1 at 8 n.13.) During the hearing, the Court directed Dr. Yang to other reports on the same website that show different numbers. (Dkt. 170-2 at 48–50.) He could not explain which report was correct. (Id.) So, by way of example, Dr. Yang testified that he relied on reports showing Douglas county had a 53 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 54 of 78 voter turnout rate of 78% in the November 2016 Election. (Id. at 39–42; see also Dkt. 149-1 at 17.)19 But, the Court directed him to a different report on the same website that showed the county had a turnout rate of 63% in the November 2016 Election. (Dkt. 170-2 at 48–50.)20 Dr. Yang could not explain which number was more accurate. (Id.) Similarly, the report from the Secretary of State that Dr. Yang used for calculating DeKalb County’s turnout rate of 76% is also in conflict with another report on the Secretary of State’s website which shows a turnout rate of 60% in that county. (Dkt. 149-1 at 14.) The same issue would apply to Dr. Yang’s calculation of the turnout rate in Henry County, the report he used showing 79.8% turnout on a countywide basis and the other showing 65.1% turnout on the same basis.21 There is thus a real concern that Dr Dr. Yang did not actually use these reports for Douglas County (since he obtained numbers directly from Douglas County itself) but he did use them for the other counties. He testified about the Douglas County report as an example of what he did for Henry County and other counties besides Douglas County. 20 The report is available at https://sos.ga.gov/index.php/elections/general_election_turnout_by_dem ographics_november_2016. 21 Compare https://sos.ga.gov/index.php/Elections/current_and_past_elections_resul ts with https://sos.ga.gov/index.php/Elections/voter_turn_out_by_demographics. 19 54 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 55 of 78 Yang substantially overstated the counties’ likely turnout rate in November. Given Dr. Yang’s prior conclusion that Henry County will experience no wait times longer than 30 minutes from the adjustment the county made, his prior assumption that November 2020 will see the same level of Election Day voting as November 2016, his failure to provide any credible (much less compelling) basis for turning away from that assumption, his decision to now use a Pew poll of national voting “preference,” and his inability to substantiate the turnout data he pulled from the Secretary’s website in the light of conflicting data located elsewhere on the same website, the Court finds Dr. Yang’s Second Report is too speculative and unreliable to show that long lines are certainly impending in Henry County for the November 2020 Election. c) Douglas County Plaintiffs’ complaint does not say that Douglas County voters previously suffered long lines as a result of voting equipment shortages. Instead, it alleges the county suffered long lines because “software errors” delayed the start of voting and rendered certain machines inoperable. (Dkt. 1 ¶¶ 176–178.) Dr. Yang nevertheless predicts that 55 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 56 of 78 14 voting locations in Douglas County will experience maximum wait times of 31–119 minutes as a result of equipment allocations. (Dkt. 1491 at 18.) But these wait times (10 of which are less than an hour) assume an extraordinarily high turnout on Election Day in November 2020. To justify that assumption, Dr. Yang did not use November 2016 Election data (as he did for Henry County in his First Report) or the 40% figure from the Pew poll (as he did for every other county in his Second Report). Instead, he used turnout projections that he received from Douglas County, which apparently obtained them from the Georgia Secretary of State’s Office. (Dkt. 170-2 at 29–30, 35–36.) Dr. Yang believes the Secretary’s numbers are the most reliable prediction of Election Day turnout because “local election officials know better about their own place.” (Id. at 86.) But he does not know how the Secretary came up with his numbers. (Id. at 86–87.) And they are wildly inconsistent with historical Election Day turnout in Douglas County, including from as recently as June 2020 (during the pandemic). Take precincts 1272 and 729, for example. Dr. Yang predicts the longest maximum wait times at those locations 56 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 57 of 78 (119 and 116 minutes). (Dkt. 149-1 at 18.) And he does so based on the assumption that 37% and 35.5% of registered voters will vote on Election Day at each location.22 But those percentages were about 22.5% and 27.2% in November 2016 (and potentially lower depending on which Secretary of State data you use).23 If Dr. Yang applied these November 2016 percentages (as he did in his First Report), his projected Election Day turnout would drop from 2,008 to 1,223 in precinct 1272, and from 1,990 to 1,526 in precinct 729. Those differences are substantial. Likewise, Dr. Yang assumes that 51.2% and 64.3% of the total turnout in 37% = 2,008 Election Day voters / 5,435 registered voters; 35.5% = 1,990 Election Day voters / 5,612 registered voters. (See Dkts. 149-1 at 18 (Dr. Yang predicting 2,008 and 1,990 Election Day voters at precincts 1272 at 729); 132-3 at 4, 7 (5,435 and 5,612 registered voters in precincts 1272 and 729).) 23 These percentages are based on data reported in the same part of the Secretary’s website on which Dr. Yang relies for other elements of his analysis: : https://results.enr.clarityelections.com/GA/63991/184321/en/selectcounty.html. But another report on the Secretary’s website shows a much lower turnout rate in Douglas County: https://sos.ga.gov/index.php/elections/general_election_turnout_by_dem ographics_november_2016 (68.6% vs. 82.9% in precinct 1272; 55.3% vs. 71.2% in precinct 729). As explained earlier in this Order, Dr. Yang did not know which data was correct. 22 57 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 58 of 78 precincts 1272 and 729 will vote on Election Day.24 But in June 2020, those percentages were just 34.8% and 44%.25 These differences are also significant. It is far from clear that Dr. Yang’s percentages make more sense than the lower ones described above. We do not know how his were calculated (beyond the fact that they were obtained from the Secretary). And they do not square with Douglas County’s actual voting history in the most recent general and pandemic-affected elections. Given these facts, the lower percentages seem at least as plausible as those used by Dr. Yang. Indeed, Dr. Yang himself would have used them (and even reduced them to account for an expected increase in absentee voting) if he were still applying the methodology on which he relied in his First Report. (See Dkt. 93-62 at 40–41.) 51.2% = 2,008 Election Day voters / 3,924 total votes cast; 64.3% = 1,990 Election Day voters / 3,095 total votes cast. (See Dkts. 149-1 at 18 (Dr. Yang predicting 2,008 and 1,990 in-person Election Day voters in precincts 1272 and 729); 132-4 at 4 (predicting total turnout of 3,924 and 3,095 voters in precincts 1272 and 729).) 25 34.8% = 599 Election Day voters / 1,720 total votes cast; 44% = 682 Election Day voters / 1,549 total votes cast. See https://results.enr.clarityelections.com/GA/Douglas/103662/web.254232/ #/summary?v=255449%2F. 24 58 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 59 of 78 These lower percentages would reduce the wait times predicted in Dr. Yang’s report. We do not know by how much, of course. Maybe it would have a big impact. Maybe not. But this uncertainty is precisely what precludes the Court from saying long lines are “certainly impending” in Douglas County. The water is muddied even further by the fact that, since receiving Dr. Yang’s report, Douglas County has added more voting equipment to all 14 locations at which Dr. Yang predicted wait times of at least 30 minutes. (Dkt. 167-1 ¶ 8.) Dr. Yang’s projections do not account for this new development, which presumably will reduce wait times even more. When all of this is combined with the Court’s broader reservations about Dr. Yang’s report, his predictions are simply too speculative to meet Plaintiffs’ burden to show a certainly impending injury. d) DeKalb County Dr. Yang predicts maximum wait times of 36 minutes, 40 minutes, and 207 minutes at three voting locations in DeKalb County. (Dkt. 1491 at 14.) DeKalb County officials have since presented evidence that they “anticipate being in a position to allocate additional poll pads and scanners” to all three locations if space and staffing permit. (Dkt. 166-1 59 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 60 of 78 ¶ 11; see id. ¶ 7 (“DeKalb VRE is in the process of ordering 50 additional scanners to be delivered for use at polling locations on Election Day. The DeKalb VRE is also working on an acquisition of 125 additional poll pads to be delivered for use at polling locations on Election Day.”).) Given this development, and the “wiggle room” already inherent in Dr. Yang’s projections, Plaintiffs have not shown long lines are certainly impending in the 36- and 40-minute locations. The 207-minute location (McWilliams) is a closer call. The projected wait time for that location assumes 40% of the total turnout in November 2020 will vote on Election Day. At his evidentiary hearing, Dr. Yang “guess[ed]” that the voting location “still might be a problem” even if he assumed (as he did in his First Report) that November 2020 will see the same Election Day turnout as November 2016. (Dkt. 170-2 at 71.)26 Although he did not “have time to do the math” during the hearing, he did do “a quick calculation” from which he determined that the maximum wait at McWilliams would likely fall from 207 minutes to 85 minutes. (Id. at 71, 82–83.) He said the “bottleneck” in that scenario He said this assumption was “likely to reduce” the 36- and 40-minute locations to less than a 30-minute wait. (Dkt. 170-2 at 72; see id. (“[Y]ou can probably do that.”).) 26 60 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 61 of 78 would be the check-in stage because there were insufficient poll pads there. (Id. at 83.) Given that DeKalb County now anticipates adding poll pads to the McWilliams location, it is unclear whether the check-in bottleneck would still arise under Dr. Yang’s model. Adding poll pads might simply move the bottleneck to the BDM stage of the process, but there is no clear evidence about that one way or the other. (See Dkt. 170-2 at 83 (testifying that, after running his “quick calculation” during the hearing, “the checkin station is the bottleneck, it’s not the voting machines”); id. at 68–69 (testifying that, under his 40% Methodology of predicting Election Day turnout at McWilliams, “if you increase the poll pads, the voters can flow through this first step very quickly, and then the supplemental BMDs could be a bottleneck” (emphasis added)).) Given the potential addition of equipment at McWilliams, Dr. Yang’s reliance on the 40% Methodology, his admittedly incomplete attempt during the hearing to calculate wait times based on Election Day turnout in November 2016, and the Court’s broader reservations about his report, Dr. Yang’s 61 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 62 of 78 predictions also fail to meet Plaintiffs’ burden to show “certainly impending” long lines at McWilliams.27 E. Plaintiffs’ Injuries Defendants Are Not Traceable to State Even if Plaintiffs had established an Article III injury based on certainly impending long lines in November, they have not shown that injury is traceable to State Defendants. element of the standing doctrine. Traceability is the second It requires “a fairly traceable connection between the plaintiff’s injury and the complained-of conduct of the defendant.” Steel Co. v. Citizens for a Better Env’t, 523 U.S. 83, 103 (1998). A “showing of proximate cause” is not necessary. Resnick v. AvMed, Inc., 693 F.3d 1317, 1324 (11th Cir. 2012). “Even a showing that a plaintiff’s injury is indirectly caused by a defendant’s actions satisfies the fairly traceable requirement.” Id. But the injury cannot “result [from] the independent action of some third party not before the court.” Hollywood Mobile Estates Ltd. v. Seminole Tribe of Fla., 641 F.3d 1259, Even assuming there were certainly impending long lines at one or two voting locations, Individual Plaintiffs have not alleged that they plan to vote there and Organizational Plaintiffs have not alleged that they will divert resources based on the risk of only fleeting long lines at those specific locations. 27 62 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 63 of 78 1265 (11th Cir. 2011). And traceability does not exist where “an independent source would have caused [plaintiff] to suffer the same injury.” Swann v. Sec’y, Georgia, 668 F.3d 1285, 1288 (11th Cir. 2012); see 13A Charles Alan Wright & Arthur R. Miller, Fed. Prac. & Proc. Juris. § 3531.5 (3d ed. Apr. 2020 Update) (“standing may be defeated by finding a different cause” and “[d]irect breaks in the causal chain have defeated standing in a wide variety of other circumstances”). Plaintiffs claim they will be injured by long lines in the November 2020 Election. Plaintiffs say those lines will be caused by (1) insufficient polling locations, (2) insufficient voting equipment allocation, (3) insufficient poll workers, (4) inadequate training of poll workers, (5) insufficient technicians, (6) insufficient time to set up polling locations, and (7) insufficient backup paper pollbooks and emergency paper ballots. As discussed above, Plaintiffs have presented future evidence only about the allocation of voting equipment. The Court finds that evidence insufficient to establish imminent injury but, even assuming it was sufficient, Plaintiffs have not shown State Defendants are responsible for the equipment misallocations they say will happen in November. 63 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 64 of 78 Granted, the Secretary of State is required by state law to provide voting equipment to the counties. O.C.G.A. § 21-2-300(a). But the counties are responsible for allocating that equipment among their precincts and for otherwise “equip[ping] polling places for use in primaries and elections.” O.C.G.A. § 21-2-70(4). The state has also “made additional equipment available to counties upon request and demonstration of need.” (Dkt. 110-5 ¶ 6; see Dkt. 159 at 59 (“[T]he counties come to the State and say here’s what we need, the State provides it.”).) And “[c]ounties additionally may procure their own equipment to the extent they need more, which the State will acceptance test and approve for use.” (Dkt. 110-5 ¶ 6.) In other words, the state provides counties with a “baseline” level of equipment and then counties must determine what to do with that equipment and whether to obtain more. (Dkt. 159 at 59 (“We provide the baseline, and then if the counties want more, they can request and/or purchase it.”); see Dkt. 171-1 ¶ 3 (the state initially provides counties with a baseline level of equipment based on past election data and the number of active voters in each county). Plaintiffs do not allege the state has violated its general duty to provide equipment. (See Dkts. 110-5 ¶ 3; 171- 64 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 65 of 78 1 ¶ 3.) So if a county does not have enough equipment in a certain precinct, the county must address the issue, not the Secretary of State or the State Election Board. Indeed, Plaintiffs conceded the Secretary of State “doesn’t have the authority to command” counties to allocate their voting equipment in a particular way — and, more generally, is not “responsible for the errors and mistakes made at the county level.” (Dkt. 159 at 105–107.)28 Plaintiffs point out that State Defendants have several significant responsibilities for elections in Georgia. And they are right. For example, the Secretary of State is Georgia’s “chief election official.” O.C.G.A. § 21-2-50(b). And State Defendants must “formulate, adopt, and promulgate . . . rules and regulations” to ensure uniformity and fairness in Georgia’s elections. O.C.G.A. § 21-2-31(1)–(2); see also id. § 21-2-31(10) (requiring State Defendants to take such “action, consistent with law, as [they] may determine to be conducive to the fair, legal, and Even Plaintiffs’ expert has concluded that, to the extent equipment shortages might arise in November, the problem will almost always be the misallocation of voting equipment within certain counties, not a shortage in the overall quantity of equipment available for deployment. How to allocate equipment within a county is a task that falls especially squarely on the counties. 28 65 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 66 of 78 orderly conduct of primaries and elections”). But, under binding caselaw, these general powers are insufficient to establish traceability. See Jacobson, 2020 WL 5289377, at *12, 14 (“voters and organizations . . . cannot rely on the Secretary’s general election authority to establish traceability,” nor can they rely on his authority “to promulgate a rule” or “issue directives”). The Court rejects Plaintiffs’ contention that the alleged injuries of which they complain are traceable to the Secretary of State simply because the Georgia Code refers to him as the “state’s chief election official.” (Dkt. 112 at 23.) Likewise, the Court rejects the notion the alleged injuries are traceable to the State Election Board simply because of its duty to ensure uniformity in the administration of election laws. (Id. at 2.) No Georgia law allows State Defendants to reach down into the county precincts and demand the relief Plaintiffs seek. (Dkt. 159 at 105–107.) Plaintiffs also rely on Grizzle v. Kemp, 634 F.3d 1314 (11th Cir. 2011), which held that Georgia’s Secretary of State was “a proper party in [an] action for injunctive and declaratory relief pursuant to Ex Parte Young.” 634 F.3d at 1316. But “Article III standing and the proper defendant under Ex parte Young are separate issues.” Jacobson, 2020 66 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 67 of 78 WL 5289377, at *13. So Grizzle does not “address[]—let alone resolve[]— the standing issues in this suit.” Id. Plaintiffs have not shown their injury (even assuming it was imminent) is traceable to the State Defendants. That means Plaintiffs lack standing to sue the State Defendants here. Plaintiffs’ claims against the State Defendants are thus dismissed for lack of subject matter jurisdiction. See Calzone v. Hawley, 866 F.3d 866, 869 (8th Cir. 2017) (“Article III standing to sue each defendant also requires a showing that each defendant caused his injury and that an order of the court against each defendant could redress the injury.” (emphasis added)); Mahon v. Ticor Title Ins. Co., 683 F.3d 59, 65 (2d Cir. 2012) (rejecting the argument that “plaintiff’s injury resulting from the conduct of one defendant should have any bearing on her Article III standing to sue other defendants”); Holland v. JPMorgan Chase Bank, N.A., 2019 WL 4054834, at *6 (S.D.N.Y. Aug. 28, 2019) (“A plaintiff proceeding against multiple defendants must establish standing as to each defendant and each claim.”). 67 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 68 of 78 III. Plaintiffs’ Requested Injunctive Relief is Unavailable Even if Plaintiffs had standing here, the Court would deny their motion for a preliminary injunction. “[A] preliminary injunction is an extraordinary and drastic remedy not to be granted unless the movant clearly establishe[s]” it is warranted. Siegel v. LePore, 234 F.3d 1163, 1176 (11th Cir. 2000); see Winter v. Nat. Res. Def. Council, Inc., 555 U.S. 7, 24 (2008) (“A preliminary injunction is an extraordinary remedy never awarded as of right.”). Plaintiffs here have not shown they are entitled to the relief they seek. On the contrary, their proposed injunction is riddled with problems that make it inappropriate.29 A. The Court Cannot Outsource the Election to Plaintiffs’ Expert Plaintiffs’ proposed injunction would order County Defendants to use Dr. Yang’s M/M/c formula “to allocate voting machines, poll pads, scanners, and technicians in a way reasonable calculated to minimize wait times” at each voting location for the November 2020 Election. (Dkt. 119-3 at 4.) More specifically, it would require Defendants to send Plaintiffs also have not shown irreparable harm for the same reasons that they fail to show imminent injury. See Alabama v. U.S. Army Corps of Engineers, 424 F.3d 1117, 1133 (11th Cir. 2005) (“[P]reventing irreparable harm in the future is the sine qua non of injunctive relief.”). 29 68 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 69 of 78 data to Dr. Yang and then implement whatever allocation plan he recommends based on that data (with the Secretary of State required to foot the bill if more equipment is needed to comply with Dr. Yang’s plan). (Id. at 4–6.) This request is truly extraordinary. And Plaintiffs cite no authority permitting (much less requiring) the Court to grant it. The United States Constitution gives states the power to set the “Times, Places and Manner of holding Elections for Senators and Representatives.” U.S. Const. art. I, § 4, cl. 1. And that power “is matched by state control over the election process for state offices.” Clingman v. Beaver, 544 U.S. 581, 586 (2005). These constitutional powers are “broad.” See Duncan v. Poythress, 657 F.2d 691, 702 (5th Cir. 1981) (“[T]he constitution leaves to the states broad power to regulate the conduct of federal and state elections.”). And courts have no business interfering with them absent a compelling need to do so. See New Georgia Project v. Raffensperger, 2020 WL 5877588, at *3 (11th Cir. Oct. 2, 2020) (Grant, J.) (the Supreme Court’s “mantra” points “in one direction— allowing the States to run their own elections”); id. at *5 (Lagoa, J. concurring) (“[F]ederal courts must be chary of hearing challenges to a 69 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 70 of 78 state’s duly enacted election procedures.”).30 When courts do interfere, they must tread carefully and provide the minimum (and least disruptive) relief necessary to redress the harm. Plaintiffs’ requested relief flies in the face of these principles. It does not require the parties to collaboratively determine how Defendants should allocate their resources for the upcoming election. Instead, it outsources the job entirely to an academic in Minnesota and requires Defendants to do whatever he says. Even Dr. Yang admits this is appropriate. He repeatedly testified that he views his proposals as data points to be considered, not blindly adopted, by election officials on the ground. (See, e.g., Dkt. 170-2 at 52– 53 (“So in terms of how to address it, it will be up to the election officials how are you going to do it. I’m not trying to say you have to use my See also Coal. for Good Governance v. Raffensperger, 2020 WL 2509092, at *4 (N.D. Ga. May 14, 2020) (courts cannot “micromanage the State’s election process”); Georgia Shift, 2020 WL 864938, at *5 (“Plaintiffs invite the Court to dictate how the Counties should properly administer their elections. The law does not allow this type of federal judicial oversight except when an election process reaches the point of patent and fundamental unfairness.”). 30 70 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 71 of 78 calculation.”).)31 That is so because there are several factors not captured by his analysis that must also be considered when deciding how to run an election. Those include the global pandemic and other “constraints” on which Dr. Yang is not an expert. (Id. at 53–54.) Dr. Yang has never even visited or seen pictures of Georgia’s voting locations. (Id. at 54.) And he has never been appointed by a court to make election decisions for a government. (Id. at 57.) He is simply not in a position to determine, single-handedly, how Defendants should run their elections here. Plaintiffs’ counsel put it best when he said at oral argument: “[W]e’ve offered experts that we’ve hired at our own expense to come up with ideas. Maybe they’re good ideas, maybe they’re bad ideas. But they’re ideas that ought to be considered.” (Dkt. 159 at 110.) Plaintiffs may be right. Perhaps Dr. Yang’s ideas should be considered; indeed, several County Defendants have now considered them. But that is not what Plaintiffs request here. They seek an injunction requiring Defendants to adopt Dr. Yang’s ideas wholesale. Plaintiffs have not shown that relief is appropriate. The Court cannot strip state officials of (See also Dkt. 170-2 at 27 (“Usually, the local election officials should know better about this.”), 28 (“I respect the counties election officials’ opinion and knowledge on their local places.”).) 31 71 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 72 of 78 their constitutionally enshrined authority over elections and reassign that authority to an academic instead (at least on the facts here). Cf. Coal. for Good Governance, 2020 WL 2509092, at *4 (dismissing case where “[t]he relief Plaintiffs seek bears little resemblance to the type of relief plaintiffs typically seek in election cases aimed to redress state wrongs”). B. Plaintiffs’ Requested Relief is Too Vague 1. Rule 65’s Specificity Requirements Other elements of Plaintiffs’ requested relief are improper for a different reason: they are too vague to enforce. Rule 65 of the Federal Rules of Civil Procedure requires an injunction to “state its terms specifically” and to “describe in reasonable detail . . . the act or acts restrained or required.” Fed. R. Civ. P. 65(d)(1). These are “no mere technical requirements.” Schmidt v. Lessard, 414 U.S. 473, 476 (1974). “The Rule was designed to prevent uncertainty and confusion on the part of those faced with injunctive orders, and to avoid the possible founding of a contempt citation on a decree too vague to be understood.” Id. “Since an injunctive order prohibits conduct under threat of judicial 72 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 73 of 78 punishment, basic fairness requires that those enjoined receive explicit notice of precisely what conduct is outlawed.” Id. The injunction must be “very explicit” and “leave[] no uncertainty in the minds of those to whom it is addressed, who must be able to ascertain from the four corners of the order precisely what acts are forbidden.” Planetary Motion, Inc. v. Techsplosion, Inc., 261 F.3d 1188, 1203 (11th Cir. 2001); 11A Charles Alan Wright & Arthur R. Miller, Fed. Prac. & Proc. Civ. § 2955 (3d ed. Apr. 2020 Update). That is, the injunction “should be phrased in terms of objective actions” and “should clearly let defendant know what he is ordered to do or not to do.” Planetary Motion, 261 F.3d at 1203; see Hughey v. JMS Dev. Corp., 78 F.3d 1523, 1531 (11th Cir. 1996) (“[A]n ordinary person reading the court’s order should be able to ascertain from the document itself exactly what conduct is proscribed.”). If “the wording of the injunction could lend itself to alternate interpretations,” it violates Rule 65’s specificity requirements. Planetary Motion, 261 F.3d at 1205. 2. Discussion Plaintiffs seek an injunction requiring Defendants to (1) provide “sufficient” secure ballot boxes, (2) ensure poll workers are “adequately” 73 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 74 of 78 trained, (3) provide “sufficient” technicians, and (4) ensure voting equipment is “adequately” tested within a “reasonable” time before each voting location opens. Such an injunction would violate Rule 65. Georgia Shift makes that clear. The plaintiffs there sought an order requiring defendants to (among other things) “provide enough polling places,” to “provide enough functioning machines at polling places,” and to “hire and train sufficient elections staff to prevent voters from having to wait in unreasonably long lines.” 2020 WL 864938, at *2. The court declined to issue the injunction because it “would be too amorphous to be capable of enforcement,” in violation of Rule 65’s specificity requirements. 2020 WL 864938, at *5. The court said the requested relief was so vague that it essentially told defendants to “obey the law.” Id. Our case is similar on both the facts and the relief requested. As in Georgia Shift, Plaintiffs here seek to compel Defendants to take election administration measures that are “adequate,” “sufficient,” and “reasonable.” But what counts as adequate, sufficient, and reasonable is subject to “alternate interpretations,” meaning Defendants could easily “misapprehend what conduct is proscribed.” Planetary Motion, 261 F.3d 74 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 75 of 78 at 1205; see Payne v. Travenol Labs., Inc., 565 F.2d 895, 898 (5th Cir. 1978) (“The word ‘discriminating,’ like the word ‘monopolizing’ . . . , is too general” to satisfy Rule 65); Brandner v. Abbott Labs., Inc., 2012 WL 27696, at *4 (E.D. La. Jan. 5, 2012) (finding that an injunction containing the word “sufficient” violated Rule 65 because “the vagueness of the term . . . places an unfair burden of interpretation on [defendant]”). This precludes Plaintiffs’ requested relief. C. Plaintiffs’ Remaining Relief is Unnecessary The final elements of Plaintiffs’ proposed injunction would order (1) Defendants to provide sufficient emergency paper ballots and paper pollbooks; (2) State Defendants to “enact a policy” requiring these paper supplies to be used when lines at are least 30 minutes long; (3) Defendants to “enact a policy” requiring poll workers to record wait times every 30 minutes; and (4) the Secretary of State to provide one technician for every ten polling locations. (See Dkt. 119-3.) Plaintiffs have not shown any of these actions are necessary to avoid long lines in November. Starting with the first item of requested relief, another court has already ordered Defendants to provide sufficient emergency paper 75 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 76 of 78 supplies to each voting location. Curling, 2020 WL 5757809, at *25. So nothing is gained by issuing another injunction requiring the same thing. (See Dkt. 159 at 113 (“[T]he backup emergency paper ballots and the paper pollbooks have been addressed in the Curling decision.”).) As for the remaining items, “an injunction is to be narrowly tailored to remedy the specific action which gives rise to it.” Valley v. Rapides Par. Sch. Bd., 646 F.2d 925, 942 (5th Cir. 1981); see Califano v. Yamasaki, 442 U.S. 682, 702 (1979) (“[T]he scope of injunctive relief is dictated by the extent of the violation established.”). Plaintiffs have not shown Defendants will cause long lines by failing to do the remaining three things they seek to compel. On the contrary, Georgia law already says poll workers may use emergency paper supplies if lines exceed 30 minutes. Ga. State Election Board Rule 183-1-12-.11(2). And the state is “increasing the number of Election Day technicians available, aiming to have a contracted technician available for each polling place.” (Dkt. 110-5 ¶ 5 (emphasis added).) The counties also have secured additional technicians of their own. Plaintiffs have not shown more is required. See Califano, 442 U.S. at 702 (“[I]njunctive relief should be no 76 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 77 of 78 more burdensome to the defendant than necessary to provide complete relief to the plaintiffs.”).32 The bottom line is that Plaintiffs lack standing because they have not shown long lines are certainly impending in November. And, even if Plaintiffs had standing, the Court cannot issue the injunction they seek because it requests relief that is either inappropriate or unnecessary. For these reasons, Defendants’ motions to dismiss are granted, Plaintiffs’ motion for a preliminary injunction is denied, and this case is dismissed for lack of subject matter jurisdiction. IV. Conclusion The Court GRANTS County Defendants’ Motion to Dismiss (Dkt. 105), GRANTS State Defendants’ Motion to Dismiss (Dkt. 106), and DENIES Plaintiffs’ Motion for Preliminary Injunction (Dkt. 92). The Court DISMISSES this action for lack of subject matter jurisdiction. It is unclear whether the Court even could order State Defendants to “enact a policy” requiring the actions Plaintiffs seek. See Jacobson, 2020 WL 5289377, at *14 (“it is doubtful that a federal court would have authority to” issue “an injunction ordering the Secretary to promulgate a rule” because “such relief would . . . raise[] serious federalism concerns”); id. (“[T]he Ex parte Young exception to sovereign immunity is limited to the precise situation in which a federal court commands a state official to do nothing more than refrain from violating federal law.”). 32 77 Case 1:20-cv-03263-MLB Document 174 Filed 10/13/20 Page 78 of 78 SO ORDERED this 13th day of October, 2020. 78