Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 1 of 44 UNITED STATES DISTRICT COURT FOR THE DISTRICT OF MARYLAND LA UNIÓN DEL PUEBLO ENTERO, et al., Plaintiffs, v. Civil Action No. 8:19-cv-02710-PX DONALD J. TRUMP, sued in his official capacity as President of the United States, et al., Defendants. MEMORANDUM OF LAW IN SUPPORT OF PLAINTIFFS’ EMERGENCY MOTION FOR TEMPORARY RESTRAINING ORDER AND/OR PRELIMINARY INJUNCTION Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 2 of 44 TABLE OF CONTENTS Page INTRODUCTION ...........................................................................................................................1 BACKGROUND .............................................................................................................................3 A. Overview of Census and Apportionment .....................................................3 B. Defendants’ Efforts to Manipulate the Census and Apportionment Process .......................................................................................................10 LEGAL STANDARD ....................................................................................................................21 ARGUMENT .................................................................................................................................21 I. II. III. Plaintiffs Are Likely to Succeed in Establishing that the Truncation of Census Operations Violates the Enumeration Clause ............................................21 A. The Bureau Cannot Complete an Accurate Count on the Current Timeline, And Plaintiffs And The Communities They Represent Will Be Undercounted ...............................................................................22 B. Response Rates and Undercounting...........................................................27 C. Precedent Demonstrates That Plaintiffs Are Entitled to Relief .................30 Truncation of Census Operations Will Cause Irreparable Harm to Plaintiffs .......34 A. Loss of Political Representation ................................................................35 B. Loss of Public Resources ...........................................................................36 The Balance of Equities and the Public Interest Support a Temporary Restraining Order or Preliminary Injunction .........................................................38 CONCLUSION ..............................................................................................................................39 i Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 3 of 44 TABLE OF AUTHORITIES Page(s) Federal Cases California v. Ross, 358 F. Supp. 3d 965 (N.D. Cal. 2019) .....................................................................................12 Carey v. Klutznick, 637 F. 2d 834 (2d Cir. 1980)....................................................................................................34 City of New York v. U.S. Dep’t of Commerce, 713 F. Supp. 48 (E.D.N.Y. 1989) ............................................................................................34 Dep't of Commerce v. New York, 139 S. Ct. 2551 (2019) ...........................................................................................11, 12, 13, 35 Franklin v. Massachusetts, 505 U.S. 788 (1992) .................................................................................................................33 J.O.P. v. U.S. Dep’t of Homeland Sec., 409 F. Supp. 3d 367 (D. Md. 2019) .........................................................................................38 Kravitz v. Dep’t of Commerce, 366 F. Supp. 3d 681 (D. Md. 2019) .......................................................................12, 36, 38, 39 La Unión del Pueblo Entero v. Ross, 353 F. Supp. 3d 381 (D. Md. 2018) .........................................................................................31 Maages Auditorium v. Prince George’s Cty., Md., 4 F. Supp. 3d 752 (D. Md. 2014) .............................................................................................21 New York v. United States Dep’t of Commerce, 351 F. Supp. 3d 502 (S.D.N.Y.).........................................................................................11, 12 Nken v. Holder,, 556 U.S. 418 (2009) .................................................................................................................38 Perez v. Texas, 891 F. Supp. 2d 808 (W.D. Tex. 2012)....................................................................................36 Reynolds v. Sims, 377 U.S. 533 (1964) .................................................................................................................35 Winter v. Nat. Res. Def. Council, Inc., 555 U.S. 7 (2008) .....................................................................................................................21 ii Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 4 of 44 Wisconsin v. City of New York, 517 U.S. 1 (1996) .........................................................................................................30, 31, 32 Federal Statutes 2 U.S.C. § 2a(a)................................................................................................................................3 13 U.S.C. § 141 ..........................................................................................................................3, 33 28 U.S.C. § 2284(a) .........................................................................................................................1 State Statutes Fla. Stat. § 11.031(1)......................................................................................................................36 Regulations 85 Fed. Reg. 44,679 (July 23, 2020) ..............................................................................................15 Executive Order on Collecting Information about Citizenship Status in Connection with the Decennial Census (July 11, 2019) (EO 13880) ................................13, 14 Other Authorities Ariz. Const. art. 4, Pt. 2, § 1 (3) .....................................................................................................36 Cal. Const. art. 21, § 1 ...................................................................................................................36 N.J. Const. art. 4, § 2......................................................................................................................36 Tex. Const. art. 3, § 26 ...................................................................................................................36 U.S. Const. amend. XIV, § 2 .........................................................................................................15 U.S. Const. art. I, § 2, cl. 3 ...................................................................................................3, 15, 30 iii Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 5 of 44 INTRODUCTION This motion seeks a temporary restraining order and/or preliminary injunction to avert imminent and irreparable harm to Plaintiffs resulting from Defendants’ plan to cease Census field operations and self-response on September 30, 2020.1 Plaintiffs are concurrently filing a Motion for Expedited Briefing Schedule and Hearing on Emergency Motion for Temporary Restraining Order and/or Preliminary Injunction. The motions can be adjudicated by a single judge because neither challenges “the constitutionality of the apportionment of congressional districts or the apportionment of any statewide legislative body” or is otherwise required to be heard by a three-judge panel. 28 U.S.C. § 2284(a).2 On August 3, 2020, Defendants announced their latest effort to manipulate the Census to undercount and deny political representation to Latinos, Asian Americans, and non-U.S. citizens. Defendants declared on that date that the time allotted to complete the 2020 Census would be truncated, with an overall timeframe four months shorter than Defendants themselves previously insisted was necessary “to ensure the completeness and accuracy of the 2020 Census.”3 This was only Defendants’ latest attempt to exclude disfavored groups from the Census and political representation, including through an unsuccessful effort to add a citizenship question to the 2020 1 Plaintiffs understand that Defendants may be planning to cease field operations at the local level even earlier, on September 16, 2020, heightening the urgency of Plaintiffs’ request on this motion. See 2020 Census Complete Count Stakeholder Working Group, YouTube, at 16:0816:46 (Aug. 19, 2020), https://www.youtube.com/watch?v=Q8fnmpyINoQ&feature=youtu.be. 2 In similar litigation in the Northern District of California, likewise involving multiple claims related to the census process, although claims directly challenging the constitutionality of apportionment have been assigned to a three judge panel including the Honorable Lucy H. Koh (see No. 20-cv-05167 (N.D. Cal. 2020)), the claims related to truncation of data collection and processing will be heard by Judge Koh alone (see No. 20-cv-0577, Dkt. 1 (N.D. Cal. 2020)). 3 U.S. Census Bureau, Statement on 2020 Census Operational Adjustments Due to COVID-19 (Apr. 13, 2020), https://www.census.gov/newsroom/press-releases/2020/statementcovid-19-2020.html. 1 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 6 of 44 Census and ongoing efforts (including those at issue here) to produce a block level citizenship dataset and exclude undocumented immigrants from congressional apportionment. Defendants’ efforts are extensive, relentless, and fast-moving. Plaintiffs, Asian American and Latino voters and organizations working on their behalf, therefore seek a temporary restraining order and/or preliminary injunction barring Defendants from proceeding with their plan to halt census field collection and self-response and deliver final population totals to the President before completing the “actual Enumeration” required by the Constitution.4 Plaintiffs satisfy all of the prerequisites for relief. Plaintiffs are likely to prevail on the merits of their Enumeration Clause claim because Defendants’ truncated operational plans will violate Defendants’ constitutional duty to conduct an “actual Enumeration” of all persons living in the United States. As numerous authorities and Defendants’ own recent public statements make clear, the rushed, half-baked process Defendants propose would not produce an actual enumeration, in part because it would significantly undercount members of Plaintiffs’ communities. Such an undercount would cause Plaintiffs irreparable harm, including the loss of political representation at multiple levels and funding for public services. Finally, because the public interest and the balance of equities favor the “actual Enumeration” the Constitution requires, a temporary restraining order and/or preliminary injunction should be entered barring Plaintiffs from proceeding with their plan to cease field collection, self-response, and data processing before completing the process Defendants themselves have identified as necessary. 4 This motion addresses the planned truncation of census operations, which presents an imminent threat of irreparable harm. By subsequent motion, Plaintiffs intend to seek further injunctive relief on their claims related to Defendants’ attempt to subtract individuals from Congressional apportionment and to create a block level citizenship dataset for redistricting. 2 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 7 of 44 Each day that passes without relief increases the likelihood that the Census will fail to achieve the required actual enumeration—a harm that, once complete, cannot be remedied. BACKGROUND A. Overview of Census and Apportionment The U.S. Constitution requires that an “actual Enumeration” of all persons living in the United States be conducted every ten years. U.S. Const. art. I, § 2, cl. 3. In the Census Act of 1976, Congress charged the Secretary of Commerce with carrying out that “decennial census of population.” 13 U.S.C. § 141(a). This count is used to determine, among other things, the number of congressional seats assigned to each state. U.S. Const. art. I, § 2, cl. 3. Under an accompanying statutory framework, the Secretary of Commerce reports to the President, by January 1 of the year following the Census (here, January 1, 2021), the “tabulation of total population by States” “as required for the appointment of Representatives in Congress.” 13 U.S.C. § 141(b). The President must transmit to Congress “the whole number of persons in each State ... as ascertained under the ... decennial census of the population” and “the number of Representatives to which each state would be entitled.” 2 U.S.C. § 2a(a). Since the adoption of the Census Act of 1976, the Census Bureau has conducted the Census four times. Each census has begun with the mailing of self-response forms to every household in America.5 Many households participate in the Census by returning these forms. 5 U.S. Census Bureau, 1980 Census of Population and Housing, Chapter 5, https://www.census.gov/prod/www/decennial.html; U.S. Census Bureau, 1990 Census of Population and Housing, Chapter 6, https://www2.census.gov/prod2/cen1990/cph-r/cph-r-2.pdf; U.S. Census Bureau, Census 2000 Operational Plan (December 2000), https://www.census.gov/ dmd/www/pdf/Operational2000.pdf; U.S. Census Bureau, 2010 Census Planning Memoranda Series: 2010 Census Nonresponse Followup Operations Assessment Report (April 30, 2012), https://www.census.gov/content/dam/Census/library/publications/2012/dec/2010_cpex_190.pdf. 3 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 8 of 44 But not every household does so; in 2010, for example, nearly one in four households did not.6 Accordingly, to complete the “actual Enumeration” the Constitution requires, the Census Bureau must conduct further field collection operations via a network of field offices. As recently explained by the Department of Commerce Inspector General, these operations are “critical to the success of a decennial census” because they “represent a final opportunity for the Bureau to reach households who have not responded, including historically hard-to-count communities.”7 In the past, the Census has included extensive field operations designed to reach every household that did not return a self-response form through nonresponse follow-up. These have included door-knocking to obtain responses from nonresponsive households, re-interviews to correct discrepancies, checks of vacant housing units, and residual canvassing referred to as nonresponse follow-up (“NRFU”) operations.8 Declaration of John Thompson, Ex. A, (“Thompson Decl.”) ¶ 14. These operations are deliberate, thorough, and time-consuming: every Census conducted in the past forty years has had a field operation lasting at least three months, and one had a field operation lasting more than six months.9 As former Census Bureau Director 6 Press Release, U.S. Census Bureau, Nation Achieves 74 Percent Final Mail Participation in 2010 Census (Oct. 21, 2010), https://www.census.gov/newsroom/releases/archives/2010_census/cb10-cn81.html. 7 U.S. Dep’t of Commerce, Office of Inspector General, 2020 Census Alert: The Census Bureau Faces Challenges in Accelerating Hiring and Minimizing Attrition Rates for Abbreviated 2020 Census Field Operations (Aug. 18, 2020), https://www.oig.doc.gov/OIGPublications/OIG20-041-M.pdf. 8 See U.S. Census Bureau, Press Release, Census Bureau Activates Additional Operations to Ensure Accurate 2020 Census Count (June 30, 2020), https://www.census.gov/newsroom/press-releases/2020/2020-additional-operations.html (describing planned field operations for the 2020 Census). 9 1980 Census of Population and Housing, Chapter 5at 26, 38 (198 day field operation in 1980); 1990 Census of Population and Housing, Chapter 6 at 34, 38 (89 day field operation in 1990); 2000 Census of Population and Housing, Chapter 5 at 250, available at https://www.census.gov/history/pdf/Census2000v1.pdf (133 day field operation in 2000); 2010 4 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 9 of 44 John Thompson has testified, and Defendants have confirmed, NRFU is a vital and indispensable step in ensuring that “hard to count” individuals, households, and communities—including Latinos, Asian Americans, and immigrants—are actually counted in the Census.10 Once data has been collected in the field and all NRFU steps have been completed, the Census Bureau must process that data to ensure it generates an accurate count of the actual persons living in the United States—a stage known as post-data collection processing or “postprocessing.” At this stage, the Bureau combines the voluminous data gathered through field operations and analyzes it to ensure that all persons are counted and that no person is counted twice. Declaration of Howard Hogan, Ex. B, (“Hogan Decl.”) ¶¶ 8, 10, 13. This too is an extensive process involving discrete steps, including checking returned questionnaires for completeness; merging self-responses received online or via paper questionnaires or the telephone; de-duplicating multiple responses from persons who are at risk of being counted twice based on their presence at both a primary residence and a vacation home, or at both their parental residence and college housing; supplementing response data with administrative records for households where NRFU was unsuccessful; imputing missing values from completed census questionnaires, including where the number of residents was not previously documented; and conducting multiple tabulations and review of aggregate results by Census Bureau demographers, among others. Hogan Decl. ¶ 13. Post-processing, like the field collection process before it, takes time to do correctly and must happen in a particular order. Hogan Decl. ¶ Census Nonresponse Followup Operations Assessment Report at 47, 108, 153, 192 (112 day field operation in 2010). 10 See Thompson Decl. ¶ 14 (“[H]ard-to-count populations are disproportionately represented in the nonresponse universe. … I view a successful NRFU as one of the most important census operations to ensuring a fair and accurate count.”); U.S. Census Bureau, 2020 Census Operational Plan—Version 4.0 (December 2018) at 212 (“The NRFU Operation is entirely about hard-to count populations.”). 5 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 10 of 44 14. The last few decennial Censuses, conducted in the past 30 years, have included a postprocessing period of four to six months.11 Both NRFU and post-processing are necessary for an “actual enumeration” of the population. Prior to August 3, 2020, the Operational Plan designed and implemented by the career staff of the Census Bureau for the 2020 Census was broadly consistent with the timelines the Census Bureau has long recognized as necessary to satisfy its constitutional obligations. Under an operational plan released in July 2019, the Bureau planned on a NRFU period of approximately three months (from mid-May to mid-August 2020), allowing for a data processing period of approximately four and a half months (from mid-August 2020 to the end of December).12 In response to the COVID-19 pandemic, Secretary of Commerce Wilbur L. Ross, Jr. and Census Bureau Director Steven Dillingham suspended field operations in March 2020. When the Bureau re-opened Area Census Offices in May and June 2020, it proposed maintaining NRFU and data processing periods of similar lengths, ultimately announcing a schedule in which primary NRFU field operations ran for approximately three months (from August 11 to the end of October), allowing for a data processing period of approximately six months (from the end of October to the end of April).13 11 1990 Census of Population and Housing, Chapter 1 at 3, Chapter 6 at 38 (155 day postprocessing period in 1990); 2000 Census of Population and Housing, Chapter 1 at 5, Chapter 5 at 250 (112 day post-processing period in 2000); 2010 Census Nonresponse Followup Operations Assessment Report at 192 and About the 2010 Decennial Census, https://www.census.gov/programs-surveys/decennial-census/decade/2010/about-2010.html (119 day post-processing period in 2010). 12 U.S. Census Bureau, 2002 Census Detailed Operational Plan for: 18. Nonresponse Followup Operation (NRFU) at 2 (July 15, 2019), https://www2.census.gov/programssurveys/decennial/2020/program-management/planning-docs/NRFU-detailed-operationalplan_v20.pdf. 13 See U.S. Census Bureau, Statement on 2020 Census Operational Adjustments Due to COVID-19 (Apr. 13, 2020), https://www.census.gov/newsroom/press-releases/2020/statement- 6 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 11 of 44 To accommodate this revised timeline, the Bureau sought approval from Congress to delay its reporting of the “tabulation of total population by States” to the President by four months.14 At the time, Defendants publicly committed themselves to the necessity of the extended timeframe. In a joint statement, Secretary Ross and Bureau Director Dillingham explained that the extension was necessary “[i]n order to ensure the completeness and accuracy of the 2020 Census.”15 President Trump said of the extension: “I don’t know that you even have to ask [Congress]. This is called an act of God. This is called a situation that has to be.”16 These statements by Defendants were consistent with those of the Census Bureau’s senior career staff, who historically have been responsible for the design and implementation of the Census.17 In late May, Census Bureau Associate Director for Field Operations Tim Olson, who is responsible for overseeing NRFU operations, stated that completing the required enumeration by the end of 2020 (as Defendants’ current plan requires) would be impossible: “We have passed the point where we could even meet the current legislative requirement of December 31. We covid-19-2020.html; U.S. Census Bureau, Updates to 2020 Census Operations (June 12, 2020), https://www.census.gov/newsroom/press-releases/2020/update-census-operations.html. 14 See U.S. Census Bureau, Statement on 2020 Census Operational Adjustments Due to COVID-19 (Apr. 13, 2020), https://www.census.gov/newsroom/press-releases/2020/statementcovid-19-2020.html. 15 Id. 16 Hansi Lo Wang, Trump Officials Ask to Delay Census Data for Voting Districts, House Seats, NPR (Apr. 13, 2020), https://www.npr.org/2020/04/13/833546675/trump-officials-ask-todelay-census-data-for-voting-districts-house-seats. 17 U.S. Census Bureau, 2020 Census Operational Plan—Version 4.0 (December 2018) at 3, https://www2.census.gov/programs-surveys/decennial/2020/program-management/planningdocs/2020-oper-plan4.pdf (“The development of the 2020 Census Operational Plan is led by the Decennial Census Management Division (DCMD), in particular a set of Assistant Division Chiefs responsible for the 2020 Census operations.”). 7 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 12 of 44 can’t do that anymore.”18 Associate Director for Decennial Census Programs Al Fontenot, who oversaw the development of the 2020 decennial Census operational plan,19 echoed this point a few months later, informing the press on July 8 that, “We are past the window of being able to get those counts by those dates at this point.”20 Accordingly, for nearly four months, Defendants and their officers proclaimed that counting efforts would continue into the fall, and postprocessing would continue until April 2021. Census Bureau staff continued to carry out their operations according to the that schedule. The public, including Organizational Plaintiffs—who are critical to reaching hard-to-count Asian American and Latino populations—relied on those representations.21 18 Hansi Lo Wong, “We’re Running Out Of Time”: Census Turns to Congress to Push Deadlines, NPR (May 27, 2020), https://www.npr.org/sections/coronavirus-liveupdates/2020/05/27/863290458/we-re-running-out-of-time-census-turns-to-congress-to-pushdeadlines. 19 See 2020 Census Operational Plan—Version 4.0 at 3 (Decennial Census Management Division responsible for design of 2020 Census operational plan); U.S. Census Bureau, Organizational Chart (Oct. 1, 2018), https://www2.census.gov/about/orgchart.pdf (DCMD is supervised by the Associate Director for Decennial Census Programs). 20 Transcript of Operational Press Briefing – 2020 Census Update, at 21 (July 8, 2020), https://www.documentcloud.org/documents/7007573-July-8-2020-Census-Bureau-Transcript-of2020.html#document/p21/a573697; see also Hansi Lo Wong, Republicans Signal They Are Willing To Cut Census Counting Short, NPR (July 28, 2020), https://www.npr.org/2020/07/28/895744449/republicans-signal-theyre-willing-to-cut-shortcensus-counting. 21 Organizational Plaintiffs understand that some of their members mistakenly believe they have additional time to complete the census; many members are hard to reach, and such mistaken reliance cannot easily be cured. See Declaration of Deborah Chen, Ex. I (“Chen Decl.”) ¶ 20. Organizational Plaintiffs have also faced unprecedented challenges to their crucial outreach efforts because of the pandemic and the truncated timeline. See Declaration of John Park, Ex. G (“Park Decl.”) ¶¶ 20-22; Declaration of Petra Falcon (“Falcon Decl.”) Ex. L, ¶¶ 11-16; Declaration of Frances Valdez, Ex. H (“Valdez Decl.”) ¶¶ 10-17; Declaration of Juanita ValdezCox, (“Valdez-Cox Decl.”), Ex. M, ¶¶ 24-27. 8 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 13 of 44 A few weeks ago, however, Defendants’ public posture on the time period necessary “to ensure the completeness and accuracy of the 2020 Census” changed dramatically. On August 3, 2020, Defendants abandoned the timeline they previously had advertised as “a situation that has to be” in favor of an approach better suited to their goal of manipulating the Census and congressional apportionment process to, in the words of an advisor central to that effort, the late Dr. Thomas Hofeller, benefit “non-Hispanic whites.”22 Director Dillingham announced that the Bureau would dramatically shorten both its field data collection and self-response periods, moving the cessation date for both of these stages from October 31 to September 30.23 This shift came only a week before field operations were scheduled to begin. Director Dillingham further stated that post-processing would be truncated by four months, from April 30, 2021 to December 31, 2020,24 despite statements by Defendants and their officers that an actual enumeration on this timeline was impossible. Census officials are unable to explain this sudden and dramatic change. When asked why the Trump administration now seeks to wrap up the Census quickly and report apportionment data this year, Bureau Director Dillingham told members of Congress, “I’m not aware of all the many reasons except to say that the Census Bureau and others really want us to proceed as rapidly as possible and to get this—get a complete and accurate count as soon as possible.”25 22 See “The Use of Citizen Voting Age Population in Redistricting,” at 9, https://www.commoncause.org/wp-content/uploads/2019/05/2015-Hofeller-Study.pdf. 23 See Statement from U.S. Census Bureau Director Steven Dillingham: Delivering a Complete and Accurate 2020 Census Count (Aug. 3, 2020), https://www.census.gov/newsroom/ press-releases/2020/delivering-complete-accurate-count.html. 24 Id. 25 See U.S. House Oversight Committee, Counting Every Person: Safeguarding the 2020 Census Against the Admin’s Unconstitutional Attacks, https://www.youtube.com/watch?v=SKXS8e1Ew7c&feature=youtu.be&t=13909 at 3:51:49. 9 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 14 of 44 B. Defendants’ Efforts to Manipulate the Census and Apportionment Process For years, even as career Census Bureau staff worked to carry out a constitutional Census, Defendants have engaged in a series of policy changes aimed at manipulating the Census and congressional apportionment process to deny political representation to members of Plaintiffs’ communities, including Latinos, Asian Americans, and immigrants. These efforts have included four discrete policy changes, each trampling constitutional or statutory commands in service of Defendants’ discriminatory agenda. The abrupt truncation of data collection efforts and elimination of key data processing steps is merely the latest step in that long-standing effort. 1. Attempted Addition of Citizenship Question to the Census On March 26, 2018, Secretary Ross issued a decisional memorandum directing the Census Bureau to add a citizenship question to the 2020 Census and to use federal and state administrative records to validate census responses to the citizenship question despite being told by the Census Bureau that doing so would negatively impact Latinos and non-U.S. citizens. The idea to include a citizenship question in the 2020 Census emerged from a study prepared in August 2015 by Dr. Thomas B. Hofeller, a leading Republican redistricting strategist and map-drawing expert for the Republican National Committee. The study, titled “The Use of Citizen Voting Age Population in Redistricting,” concluded that “[a] switch to the use of citizen voting age population as the redistricting population base”—excluding non-U.S. citizens from apportionment—“would be advantageous to Republicans and non-Hispanic whites.”26 Dr. Hofeller acknowledged in his memo that the shift from redistricting based on total population to citizen voting-age population was a “radical departure”— and one that would alienate Latino 26 See “The Use of Citizen Voting Age Population in Redistricting,” at 9, https://www.commoncause.org/wp-content/uploads/2019/05/2015-Hofeller-Study.pdf. 10 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 15 of 44 voters. Dr. Hofeller’s study also explained that this switch would be “functionally unworkable” without “add[ing] a citizenship question to the 2020 Decennial Census form”27 After President Trump’s election and inauguration, Defendants began to act on Dr. Hofeller’s blueprint for consolidating political power through racial discrimination. As early as the presidential transition, Trump administration officials conferred with then-Secretary of State of Kansas Kris Kobach regarding the citizenship question.28 In a July 14, 2017 email, Mr. Kobach reconnected with Secretary Ross to discuss the 2020 Census, writing that “it is essential” that the citizenship question be added to the 2020 Census in order to exclude from apportionment “aliens who do not actually ‘reside’ in the United States” and offering to help “accomplish the addition of this question.”29 After consulting with “senior Administration officials” and Mr. Kobach, Secretary Ross repeatedly urged staff to take action on adding a citizenship question— which staff understood to require devising an alternative justification for adding the question that could survive judicial scrutiny.30 Secretary Ross and the Department of Commerce solicited the assistance first of the DOJ and then of Department of Homeland Security, before finally securing the participation of the DOJ and then-Attorney General Jeff Sessions.31 This collaboration 27 Id. at 4. 28 U.S. House Committee on Oversight and Reform. White House Interference Committee Interview of Kris Kobach, June 7, 2019, at 10, https://oversight.house.gov/sites/democrats. oversight.house.gov/files/2019-06-07.%20COR%20DEM%20Memo%20re%20Kris%20 Kobach%20Interview.pdf. 29 Kris Kobach, Follow up on our phone call (2017), https://apps.npr.org/documents/ document.html?id=4500011-1-18-Cv-02921-Administrative-Record#document/p776/a428457. 30 New York v. United States Dep’t of Commerce, 351 F. Supp. 3d 502, 549-553 (S.D.N.Y.), aff’d in part, rev’d in part and remanded sub nom. Dep't of Commerce v. New York, 139 S. Ct. 2551 (2019), and appeal dismissed, No. 19-212, 2019 WL 7668098 (2d Cir. Aug. 7, 2019). 31 Id. at 554-556. 11 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 16 of 44 resulted in a letter from the DOJ requesting the addition of the question to the 2020 Census for the purposes of enforcing the Voting Rights Act of 1965 (the “DOJ Letter”).32 The DOJ Letter included phrasing from Dr. Hofeller’s study and incorporated the substantive content of communications between Commerce and DOJ, as well as the head of President Trump’s transition team. Secretary Ross falsely claimed that the reason for adding the citizenship question was to collect citizenship data to assist the DOJ in its enforcement of the Voting Rights Act, although the true rationale was instead to deprive racial and ethnic minorities of political power through the census process.33 President Trump himself confirmed that, contrary to the purported goal of assisting the DOJ with enforcement of the Voting Rights Act, the “[n]umber one” reason to add a citizenship question to the Census was congressional districting.34 Several lawsuits successfully challenged this addition of the citizenship question as unlawful, and three federal courts permanently enjoined its addition to the 2020 Census. See Kravitz v. Dep’t of Commerce, 366 F. Supp. 3d 681, 755 (D. Md. 2019); California v. Ross, 358 F. Supp. 3d 965, 1050 (N.D. Cal. 2019); New York v. Dep’t of Commerce, 351 F. Supp. 3d 502, 32 Id. 33 Dep’t of Commerce v. New York, 139 S. Ct. 2551, 2575 (2019) (“Altogether, the evidence tells a story that does not match the explanation the Secretary gave for his decision. … [H]ere the VRA enforcement rationale—the sole stated reason—seems to have been contrived.”); Michael Wines, Deceased G.O.P. Strategist’s Hard Drives Reveal New Details on the Census Citizenship Question, N.Y. Times (May 30, 2019), https://www.nytimes.com/2019/05/ 30/us/census-citizenship-question-hofeller.html. 34 The White House, Remarks by President Trump Before Marine One Departure, July 5, 2019, https://www.whitehouse.gov/briefings-statements/remarks-president-trump-marine-onedeparture-51 ( “Q: What’s the reason, Mr. President, for trying to get a citizenship question on the census?; THE PRESIDENT: Well, you need it for many reasons. Number one, you need it for — ; Q: (Inaudible) reason?; THE PRESIDENT: — Congress. You need it for Congress, for districting.”). 12 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 17 of 44 679 (S.D.N.Y.), aff’d in part, rev’d in part and remanded sub nom. Dep’t of Commerce v. New York, 139 S. Ct. 2551 (2019). During the litigation, the Trump Administration stood by Secretary Ross’s pretextual rationale, denying that the addition of the citizenship question was directed at congressional apportionment as “a speculative conspiracy theory that is unsupported by the evidence.”35 On June 27, 2019, the United States Supreme Court affirmed the injunction, concluding that the “VRA enforcement rationale—the sole stated reason [for adding a citizenship question]—seems to have been contrived.” Dep’t of Commerce v. New York, 139 S. Ct. 2551, 2575 (2019). Nevertheless, the effects of Defendants’ attempts to include a citizenship question in the Census continue to reverberate, exacerbating fear, mistrust, and reluctance to respond to the Census among communities of color and immigrant communities. Chen Decl. ¶¶ 20-24; Park Decl. ¶¶ 14-16. This has compounded the difficulty of reaching hard-to-count communities and required additional efforts on the part of community stakeholders, efforts that will be thwarted by the truncated timeline. Chen Decl. ¶¶ 19-23. 2. Collection of Citizenship Data Through Other Sources Undaunted, Defendants pursued a second strategy: equipping states to use block level citizen voting-age population (CVAP) data as a basis for congressional and state redistricting. On July 11, 2019—two weeks after the Supreme Court’s decision on the citizenship question— President Trump issued an executive order requiring all executive departments and agencies to provide the Department of Commerce “maximum assistance … in determining the number of citizens and non-citizens in the country.” Executive Order on Collecting Information about Citizenship Status in Connection with the Decennial Census (EO 13880), at § 1 (July 11, 35 Letter from Noel J. Francisco, Solicitor General, to Honorable Scott S. Harris, Clerk, Supreme Court of the United States (June 25, 2019), https://www.supremecourt.gov/DocketPDF/ 18/18-966/103988/20190625110441312_Letter%20Response%2018-966.pdf. 13 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 18 of 44 2019).36 EO 13880 also instructed the Census Bureau to create an inter-agency working group to collect citizenship data in connection with the 2020 Census for redistricting and directed the Department of Commerce to obtain federal and state administrative records to support this effort. Id. at §§ 1, 3(b), 3(d). In support of these directives, EO 13880 listed several pretextual reasons for collecting citizenship data, including: (1) “to help us understand the effects of immigration on our country and to inform policymakers considering basic decisions about immigration policy”; (2) “the lack of complete data on numbers of citizens and aliens hinders the Federal Government’s ability to implement specific programs and to evaluate policy proposals for changes in those programs”; and (3) “data identifying citizens will help the Federal Government generate a more reliable count of the unauthorized alien population in the country,” which is necessary to “evaluat[e] many policy proposals.” EO 13880 at § 1. But the order also revealed the true reason for its issuance and Secretary Ross’s prior decision to add a citizenship question to the 2020 Census: so that “State and local legislative districts [can redistrict] based on the population of voter-eligible citizens.” EO 13880 at § 1. On the day he issued the order, President Trump corroborated that intent, saying that “[t]his information is also relevant to administering our elections” because “[s]ome states may want to draw state and local legislative districts based upon the voter-eligible population,” and that the Supreme Court “would not review certain types of districting decisions, which could encourage states to make such decisions based on voter eligibility.” Id. 36 Available at https://www.whitehouse.gov/presidential-actions/executive-order-collectinginformation-citizenship-status-connection-decennial-census. 14 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 19 of 44 3. Exclusion of Undocumented Persons from Apportionment Totals On July 21, 2020, President Trump issued a presidential memorandum declaring that it is “the policy of the United States to exclude from the [congressional] apportionment base aliens who are not in a lawful immigration status under the Immigration and Nationality Act.” Memorandum on Excluding Illegal Aliens from the Apportionment Base Following the 2020 Census (“Memorandum”), 85 Fed. Reg. 44,679-80 (July 23, 2020). The Memorandum reflects a radical departure from long-established policy and practice. Every prior apportionment under the current statutory framework has included the total U.S. population, including undocumented individuals, in the apportionment totals presented to Congress. The Memorandum is also facially inconsistent with the Constitution’s clear command that “the whole number of persons in each State, excluding Indians not taxed,” be counted. U.S. Const. amend. XIV, § 2 (emphasis added); see id. art. I, § 2, cl. 3. To advance this policy, the Memorandum stated that the President will exclude “aliens who are not in a lawful immigration status” from his calculation of the “whole number of persons in each State” to “settle[] the apportionment of Representatives among the States.” 85 Fed. Reg. 44,679-80 (July 23, 2020) (internal citation and quotation marks omitted). President Trump’s official statements accompanying the release of the Memorandum demonstrate that the purpose of the policy is not to advance principles of representative democracy, but to discriminate against undocumented individuals, non-U.S. citizen immigrants more generally, and U.S. citizens who live near undocumented individuals. President Trump declared that the policy was needed because “the radical left is trying to erase the existence of this concept [of U.S. citizenship] and conceal the number of illegal aliens in our country” as part of “a broader left- 15 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 20 of 44 wing effort to erode the rights of Americans citizens.” According to the President, “we should not give political power to people who should not be here at all.”37 The Presidential Memorandum has further discouraged communities of color from participating in the Census. As one expert testified, the Memorandum “will almost certainly increase the number of people unwilling to share their name.” Hogan Decl. ¶ 33; see also Park Decl. ¶ 16; Chen Decl. ¶ 24. 4. Truncation of Census Field Collection and Data Processing On August 3, 2020, Defendants’ efforts to manipulate the Census and apportionment process reached the operational plans for the 2020 Census. On that date, Census Bureau Director Dillingham announced that the Census Bureau would depart from the plan and timeframe it previously had characterized as necessary to its mission, and instead shorten by four weeks its field operations and the self-response period—providing a mere 52 days to ensure that households that had not yet responded to the Census are included in the count.38 Again, Defendants had previously budgeted 90 days for field operations, which in prior census years required as long as six months to complete.39 Thompson Decl. ¶ 23. Director Dillingham stated that the post-processing period—which is “necessary to produce accurate Apportionment 37 The White House, Statement from the President Regarding Apportionment (July 21, 2020), https://www.whitehouse.gov/briefings-statements/statement-president-regardingapportionment. 38 See Statement from U.S. Census Bureau Director Steven Dillingham: Delivering a Complete and Accurate 2020 Census Count (Aug. 3, 2020), https://www.census.gov/newsroom/ press-releases/2020/delivering-complete-accurate-count.html; U.S. Gov’t Accountability Office, 2020 Census: Recent Decision to Compress Timeframe Poses Additional Risks to an Accurate Count (Aug. 2020), at 5, https://www.gao.gov/assets/710/709015.pdf. 39 U.S. Census Bureau, 2002 Census Detailed Operational Plan for: 18. Nonresponse Followup Operation (NRFU) at 2 (July 15, 2019), https://www2.census.gov/programssurveys/decennial/2020/program-management/planning-docs/NRFU-detailed-operationalplan_v20.pdf. See infra n.8. 16 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 21 of 44 Counts,” Hogan Decl. ¶ 14—would be truncated by four months, from April 30, 2021 to December 31, 2020. This change will force the Census Bureau to complete a time-consuming multi-step post-processing procedure within a time frame two months shorter than that budgeted before the COVID-19 pandemic—“an enormous if not impossible task.” Hogan Decl. ¶ 57. The truncated census operations are a sharp reversal from prior practice and are likely to lead to a significant undercount of members of Plaintiffs’ communities, including Latinos, Asian Americans, and immigrants. NRFU field operations are essential for counting households that do not self-respond to the census questionnaire, a greater proportion of which tend to be minority households. In the 2010 Census, for example, Non-Hispanic White households self-responded via mail at a rate of 82.5 percent, whereas Hispanic households self-responded at a rate of 69.7 percent, Black households self-responded at a rate of 70 percent, Asian American households self-responded at a rate of 75.4 percent, and other communities of color self-responded at even lower rates.40 Racial and ethnic minorities also are over-represented in the census tracts that are hardest to count—meaning those with the lowest self-response rates. Nearly 66 percent of the population in the hardest to count tracts are minorities, even though they are only 39.8 percent of the population. Declaration of William O’Hare (“O’Hare Decl.”), Ex. C, ¶ 109. As the Department of Commerce Inspector General observed, NRFU “represent[s] a final opportunity for the Bureau to reach households who have not responded, including historically hard-to-count communities.”41 40 U.S. Census Bureau, 2010 Census Planning Memoranda Series: 2010 Census Mail Response/Return Rates Assessment Report at 24, 26 (June 6, 2012), https://www.census.gov/ content/dam/Census/library/publications/2012/dec/2010_cpex_198.pdf. 41 U.S. Dep’t of Commerce, Office of Inspector General, 2020 Census Alert: The Census Bureau Faces Challenges in Accelerating Hiring and Minimizing Attrition Rates for Abbreviated 2020 Census Field Operations (Aug. 18, 2020), https://www.oig.doc.gov/OIGPublications/OIG20-041-M.pdf. 17 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 22 of 44 This year, in the midst of a global pandemic and a series of natural disasters including hurricanes and wildfires, Census staff face an especially enormous—and daunting—task: to visit and attempt to enumerate approximately 56 million addresses that have not yet responded to the Census42 and to do so in the shortest field operations period in modern Census history. This is a full 8 million households more than the Census Bureau sought to enumerate through NRFU in 2010, and Census staff must do so in 7 weeks instead of 10 weeks. O’Hare Decl. ¶ 116. On top of this, the Bureau is facing a staffing shortfall of some 80,000 enumerators—more than onequarter of those the Department sought to hire for the NRFU process.43 In fact, the Department of Commerce Inspector General warned on August 18 that because of this staffing shortfall, “the Bureau is at risk of not conducting a complete and accurate 2020 Census.”44 Former-Census Bureau Director Thompson has testified that, based on his experience, the compressed timeline “will not be adequate to complete NRFU without accepting lower quality enumerations and incompletely enumerating the traditionally hard-to-count populations.” Thompson Decl. ¶ 21. These issues are compounded in immigrant communities such as those in Queens County, New York; Maricopa County, Arizona; Harris County, Texas; and the Rio Grande Valley of Texas,45 where facility in languages other than English is necessary for effective 42 U.S. Census Bureau, Door-to-Door Visits Begin Nationwide for 2020 Census (Aug. 11, 2020), https://www.census.gov/newsroom/press-releases/2020/door-to-door-visits-beginnationwide.html. 43 U.S. Dep’t of Commerce, Office of Inspector General, 2020 Census Alert: The Census Bureau Faces Challenges in Accelerating Hiring and Minimizing Attrition Rates for Abbreviated 2020 Census Field Operations (Aug. 18, 2020), https://www.oig.doc.gov/OIGPublications/OIG20-041-M.pdf. 44 Id. See also Thompson Decl. ¶ 32. 45 The Rio Grande Valley encompasses Cameron County, Hidalgo County, Starr County, and Willacy County, substantially rural areas along Texas’s southern border. Queens County is wholly subsumed within New York City. Harris County includes the City of Houston, Texas, 18 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 23 of 44 communication with many residents. Chen Decl. ¶ 18 (Organizational Plaintiff has traditionally circulated outreach materials in “Mandarin, Cantonese, Vietnamese, Spanish, Burmese, Arabic, Korean, Urdu, and Tagalog”); Park Decl. ¶ 18; Valdez Decl. ¶¶ 8-9; Falcon Decl. ¶ 7. A Census Bureau policy barring non-citizens from serving as translators has dramatically limited the applicant pool and compounded the challenges the Bureau faces in contacting hard-to-count populations. Chen Decl. ¶ 17; Park Decl. ¶ 19. Moreover, given the severely compressed timeframe for NRFU, allied stakeholders whose mission includes promoting census participation among communities of color—and whose efforts in so doing have historically been very effective—will not have the time or the resources to conduct outreach campaigns with normal efficiency or effectiveness, particularly in light of the COVID-19 pandemic. Chen Decl. ¶¶ 1921; Park Decl. ¶¶ 17-26; Valdez ¶¶ 13-17; Falcon Decl. ¶¶ 17-21. Similarly, Defendants’ elimination of four months of post-processing following the completion of NRFU strips Census Bureau professional staff of the time they need (as Defendants earlier acknowledged) to satisfy the Bureau’s duty “to count everyone once, only once, and in the right place.” Hogan Decl. ¶ 8. The Census Bureau has spoken of “[s]treamlining backend processing to deliver apportionment counts by … December 31, 2020,” but have provided no information on which steps will be rushed, or skipped entirely. Hogan Decl. ¶ 16. Post-processing is a detailed, multi-step process. Skipping steps will create unacceptable risk of serious error, preventing an actual enumeration and further exacerbating the differential undercount of Plaintiffs’ communities. Hogan Decl. ¶¶ 9, 14, 21, 38, 43, 51, 59, 69. and is the third-most populous county in the United States. Maricopa County includes the City of Phoenix, Arizona, and is the fourth-most populous county in the United States. 19 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 24 of 44 Early cessation of data collection and the compression of the data processing phase continues Defendants’ efforts to exclude minority populations from the 2020 Census count and virtually ensures that those communities that have seen low response rates to date will be undercounted. See infra Argument, I.A-B. As the Census Bureau is well aware, those communities currently include—and have historically included—Black, Latino, Asian American, and immigrants households. Indeed, racial and ethnic minority communities have had lower self-response rates than white households in the past three censuses, and these lower selfresponse rates are correlated with higher net undercount rates and omission from the final census count.46 O’Hare Decl. ¶¶ 29-91, 100. Furthermore, in 2010, and again in 2020, housing units that respond later in the Census data collection period have been disproportionately Latino, Asian American, and Black rather than Non-Hispanic White, making the end of the self-response period and NRFU (both of which have been truncated here) particularly important for counting these households. O’Hare Decl. ¶¶ 101-111. And the curtailment of the post-processing period will make it impossible for the Bureau’s professional staff to complete the intricate data cleaning and processing functions that are essential to achieve an actual enumeration, leading to an undercount that disproportionately affects areas with minority populations. Hogan Decl. ¶ 14. In short, there is a grave and immediate risk that, should the Census conclude on the Department of Commerce’s current timeline, those communities will be undercounted and, as a result, 46 Net undercount and omission rates are “both measures of Census accuracy, but they capture different parts of Census accuracy.” O’Hare Decl. ¶ 17. The net undercount is a balance between people missed (omissions), those included erroneously (those double counted and those inappropriately included in the census, like foreign tourists), and those imputed. If the number of omissions is higher than the number of erroneous inclusions and whole person imputations, there is a net undercount. Id. ¶ 18. Omissions capture the number and share of a population that are missed in the Census and are defined by the Census Bureau. In other words, “omissions are people who should have been enumerated in the United States Census but were not.” Id. ¶ 20. 20 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 25 of 44 politically underrepresented for at least the next decade. Hogan Decl. ¶¶ 21, 38, 43, 51, 59, 69; Thompson Decl. ¶¶ 3-4, 13, 21, 22, 32; O’Hare Decl. ¶¶ 95-100, 111, 121-127; Declaration of Kimball W. Brace (“Brace Decl.”), Ex. E, ¶¶ 32-33. LEGAL STANDARD Emergency injunctive relief—here, a TRO and/or preliminary injunction—is warranted where a movant demonstrates, by a preponderance of the evidence, the following elements: “(1) a likelihood of success on the merits; (2) a likelihood of suffering irreparable harm in the absence of preliminary relief; (3) that the balance of equities tips in the party’s favor; and (4) that issuing the injunction is in the public interest.” Winter v. Nat. Res. Def. Council, Inc., 555 U.S. 7, 20 (2008); Maages Auditorium v. Prince George’s Cty., Md., 4 F. Supp. 3d 752, 760 & n.1 (D. Md. 2014), aff’d, 681 F. App’x 256 (4th Cir. 2017). ARGUMENT I. Plaintiffs Are Likely to Succeed in Establishing that the Truncation of Census Operations Violates the Enumeration Clause Plaintiffs are likely to succeed on the merits of their claim that Defendants’ truncation of Census operations violates the Enumeration Clause of the Constitution. The testimony of Plaintiffs’ experts (including a former Director of the Census Bureau), publicly available data from the current Census and historical censuses, and the public statements of Defendants themselves demonstrate that communities of color, including Latinos, Asian Americans, and immigrants, are likely to be severely undercounted in the 2020 Census if Defendants are allowed to proceed with their truncated—and unconstitutional—census plan. Hogan Decl. ¶¶ 21, 38, 43, 51, 59, 69; Thompson Decl. ¶¶ 3, 13, 21, 22, 32; O’Hare Decl. ¶¶ 125-127. 21 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 26 of 44 A. The Bureau Cannot Complete an Accurate Count on the Current Timeline, And Plaintiffs And The Communities They Represent Will Be Undercounted The Census Bureau’s procedures for data collection and processing are necessarily timeintensive and labor-intensive. As explained above, accurate tabulation and verification of census data has traditionally required a minimum of three months, while post-processing has typically taken four to six months. Historically, these processes, especially NRFU, have been essential to ensuring that minority households are counted accurately. Yet Defendants, amid a pandemic, natural disasters, and staffing shortfalls, have elected to administer the shortest field operation in modern history and to cut the post-processing period perilously short, virtually guaranteeing that the total U.S. population, and particularly minority households, will be undercounted. Defendants’ own public statements, and those of their officers, confirm that, in view of the COVID-19 pandemic and its effect on the census count, an accurate count cannot be completed within the current time frame. See supra 7-8. These well-documented concerns are further confirmed by the testimony of former Census Director John Thompson. The prior operational plan for the 2020 Census estimated that conducting an adequate enumeration would require about 300,000 enumerators over the course of about 80 days. Thompson Decl. ¶¶ 16, 21. Defendants’ new plan, however, reduces that time period to 52 days—subtracting more than 30 percent of the time that the Bureau initially allocated for field operations.47 At the same time, as of August 17, the Bureau has only recruited and trained about 220,000 enumerators, meaning that the Bureau’s staffing for the NRFU phase is lagging by more than 27 percent for this vital phase of the process. Thompson Decl. ¶ 21. Moreover, approximately 15 percent of the NRFU 47 U.S. Gov’t Accountability Office, 2020 Census: Recent Decision to Compress Timeframe Poses Additional Risks to an Accurate Count (Aug. 2020), at 5, https://www.gao.gov/assets/710/709015.pdf. 22 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 27 of 44 workload is located in areas where the Bureau is falling 50 percent short of its hiring goals. Id. The Bureau is also experiencing higher-than expected attrition among enumerators. According to the U.S. Government Accountability Office (GAO), the Bureau expected that 10 percent of enumerators that started training would not show up for work in the field, but as of August 18, approximately 35 percent of employees that started training did not show up.48 To make matters worse, NRFU workloads will be relatively higher in areas with lower self-response rates—areas that contain higher proportions of Black, Latino, and immigrant populations relative to the white non-Hispanic population—making it even more difficult to obtain an accurate enumeration of these populations given Defendants’ extraordinary decision to truncate field operations. Thompson Decl. ¶ 21-22. As former Director Thompson has testified, this reduced staffing, combined with the truncated timeframe, will “not be adequate to complete NRFU without accepting lower quality enumerations and incompletely enumerating the traditionally hard-to-count populations,” leading to “increased undercounts.” Id. ¶ 21. The NRFU operation, which is the most expensive and labor-intensive phase of the Census, is challenging under any conditions. But it is likely to be even more challenging in 2020 because COVID-19 has diminished the efficacy of door-knocking, as many Americans are reluctant to interact with strangers, and families residing in dense communities or high-density housing have relocated to permit social distancing.49 See Hogan Decl. ¶ 43, 58; Park Decl. ¶¶ 20, 24-26. Also, as the GAO has noted, “a shorter NRFU operation gives these efforts less time to motivate response among hard-to-count populations.”50 The truncated NRFU process likely 48 Id. at 8. 49 Id. at 17. 50 Id. 23 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 28 of 44 will be unable to compensate for the significant undercount anticipated (and already witnessed) in communities of color. Park Decl. ¶¶ 14, 27; Chen Decl. ¶¶ 19, 25. Defendants’ abrupt curtailment of the self-response period by four weeks is also likely to produce a differential undercount of minority populations compared to white households. NonHispanic Whites have constituted a disproportionately large share of the responses early in the data collection period, while Latino (whom the Census terms “Hispanics”), Asian Americans, and Black respondents make up a disproportionately large share of the response near the end of the data collection period. O’Hare Decl. ¶¶ 101-111. The truncation of the self-response period, along with a curtailed NRFU operation, is likely to result in more Latino, Asian American, and Black households being missed, id. ¶ 111, preventing an actual enumeration of these populations. As a result of Defendants’ truncation of field operations, the Census Bureau will have to rely less on direct in-person contact attempts to reach non-responsive households and instead rely on other measures to estimate the population. Although these methods play an important role in census operations, they have serious limitations and cannot compensate for a truncated and incomplete NRFU. Over-reliance on these methods is likely to contribute to a more severe undercount of hard-to-count populations than in past recent censuses. Thompson Decl. ¶¶ 21-24; Hogan Decl. ¶¶ 9, 38, 43, 47, 50, 52, 59-60; Declaration of Amy O’Hara (“O’Hara Decl.”), Ex. D, ¶¶ 5, 11-24. For example, the Bureau may rely more heavily on proxy enumerations, which are obtained by asking people other than the residents of non-responsive households (e.g., neighbors or apartment managers) for information. Thompson Decl. ¶ 21; O’Hara Decl. ¶ 11. Proxy enumerations, however, have a high error rate. In the 2010 Census, the erroneous enumeration rate for the proxy enumeration was 6.7 percent—over twice the overall rate of 3.3 percent. Thompson Decl. ¶ 21. Moreover, truncated field operations will limit the Bureau’s 24 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 29 of 44 ability to obtain proxy responses for non-responsive housing units, making this an even less effective substitute for a complete NRFU. O’Hara Decl. ¶¶ 6(a), 11. The Bureau may turn to administrative records, such as records from the IRS, Medicare, or the Social Security Administration, to enumerate non-responsive households. O’Hara Decl. ¶¶ 12-15, 23. But these records have significant limitations. For example, individual income tax data, on which the Bureau plans to rely, will miss many individuals, including all those not in the labor force and most children. O’Hara Decl. ¶¶ 12-13, 22. Such data may be especially unreliable for a year in which IRS processing was dramatically disrupted by the pandemic. O’Hara Decl. ¶ 13. Moreover, although the Census Bureau has long planned to use administrative records to enumerate a certain portion of non-responsive households, expanded use of those records to compensate for a truncated NRFU is not supported by the research the Bureau has used to date. Thompson Decl. ¶ 21; O’Hara Decl. ¶¶ 5-6, 11, 20-24. Instead, Bureau research has shown that administrative records are not a substitute for NRFU, especially for Latinos, Asian Americans, and immigrants. Thompson Decl. ¶ 21; Hogan Decl. ¶ 52; O’Hara Decl. ¶¶ 21, 23. Thus, use of administrative records beyond the planned levels for NRFU would be even less representative of the hard-to-count populations than a complete NRFU. Thompson Decl. ¶ 21; O’Hara Decl. ¶¶ 21, 23-24. Finally, the Bureau will likely rely more heavily on a statistical technique known as “imputation” to fill in information for non-responsive households. Thompson Decl. ¶ 21. This technique imputes household size and other characteristics to non-responsive households based on information from resolved housing units. Thompson Decl. ¶ 21; O’Hara Decl. ¶ 5. Its accuracy is dependent on robust fieldwork, however, because any undercounts of resolved housing units are carried forward in the imputation and not corrected. Thompson Decl. ¶ 21; 25 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 30 of 44 O’Hara Decl. ¶¶ 13-14, 18-19. Historically, the percent of count imputations used in the Census has fluctuated but has tended to be less than 0.5 percent. Hogan Decl. ¶ 55. The number of housing units requiring whole person imputation (and thus the ratio of imputation to fieldwork) is likely to be much larger than in 2010, due to the reduction in field operations and staff limitations, increasing the likelihood of an undercount that would impact minorities disproportionately. Thompson Decl. ¶ 21; Hogan Decl. ¶¶ 56, 59-61. The truncation of field operations is not the only aspect of Defendants’ plan that will harm minority communities. The Census Bureau’s compressed timeline for post-processing, and likely shortening or elimination of several critical steps in the data verification process, will ensure that data inaccuracies will persist, making it likely that minority households will be undercounted at higher rates than other groups. Hogan Decl. ¶¶ 9, 60-61, 64-66. For example, this month the Census Bureau quietly informed states and territories that it plans to eliminate the final phase of Count Review—called Census Count and File Review—the operation that helps the Bureau spot anomalies and unexpected results in the preliminary housing unit and population counts (including group quarters) before they are finalized, while they can still be fixed.51 This is an essential step because, as Plaintiffs’ experts have testified, NRFU proxy responses are universally acknowledged to be less accurate than self-responses. Thompson Decl. ¶ 21. Beyond exacerbating the undercounting of Plaintiffs’ communities, the Bureau’s plan may necessitate limiting the process of de-duplication—that is, correcting for the “overcounting” 51 Hansi Lo Wang, Running Out of Time, Census Scales Back a Critical Step: Checking Its Own Work, NPR (Aug. 29, 2020), https://www.npr.org/2020/08/29/905846761/the-censusscales-back-a-critical-step-checking-its-own-work; Jeffrey Mervis, Will the 2020 census numbers be good enough, and how soon will we know?, SCIENCE (Aug. 24. 2020), https://www.sciencemag.org/news/2020/08/will-2020-census-numbers-be-good-enough-andhow-soon-will-we-know. 26 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 31 of 44 of certain U.S. residents. The 2010 Census saw 8.5 million duplicate responses, up from 6.6 million in the 2000 Census.52 Duplication occurs in a number of circumstances, as, for example, when people own multiple homes, or college students are counted both at school and at their parents’ homes. Hogan Decl. ¶¶ 25-48. These scenarios occur disproportionately in wealthier communities, and are less likely to occur in communities of color. See Hogan Decl. ¶¶ 38, 43. The Bureau has the ability to identify such duplicate enumerations, but it needs “time to study the problem, identify solutions and implement the corrections.” Hogan Decl. ¶ 39. In short, any efforts to skip post-process steps or rush them—which the Bureau has acknowledged will be necessary under the truncated operational plan53—is likely to result in errors and, consequently, an undercount that disproportionally affects areas with minority populations. Hogan Decl. ¶ 14. B. Response Rates and Undercounting Many of the communities in which Plaintiffs reside have already observed low response rates. See, e.g., Declaration of Gerardo E. Gonzalez (“Gonzalez Decl.”), Ex. T, ¶¶ 9-12. Latino communities in particular have historically experienced low response rates relative to other communities. See supra 20, 23. The 2020 Census is following the same pattern. As of August 24, 2020, Census tracts where Latino and Black residents are a plurality have lower self-response rates than tracts where Non-Hispanic Whites are a plurality. O’Hare Decl. ¶ 100. Furthermore, as of August 24, Asian American plurality tracts in the largest cities of metropolitan areas (over 52 U.S. Census Bureau, Census Coverage Measurement Estimation Report (May 22, 2012), at 12 tbl. 3, https://www2.census.gov/programs-surveys/decennial/2010/technicaldocumentation/methodology/g-series/g01.pdf. 53 Albert E. Fontenot, Jr. & Timothy P. Olson, Review of 2020 Operational Plan Schedule, U.S. Census Bureau, at 9 (Aug. 17, 2020), https://www.census.gov/content/dam/Census/newsroom/press-kits/2020/2020-operational-planschedule-review.pdf (noting the need to “[s]treamlin[e] backend processing to deliver apportionment counts by the statutory deadline of December 31, 2020”). 27 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 32 of 44 1 million) have a mean self-response rate of 58.8 percent, well below the national response rate of 64 percent, and the mean self-response rate for Asian American plurality tracts in rural areas (56.9 percent) is well below the national average. O’Hare Decl. ¶ 80. These trends indicate that Latino, Asian American, and Black households are on a trajectory to have higher net undercounts and omissions rates in the 2020 Census when compared to Non-Hispanic White households. O’Hare Decl. ¶¶ 77, 80, 91. The truncated NRFU period means that there will be less time available to contact and count individuals in these demographic groups with the lower selfresponse rates and ensure that they are enumerated in the 2020 Census. O’Hare Decl. ¶ 111. Several Plaintiffs reside in communities with low response rates and thus are likely to be undercounted. As just a few examples: • Plaintiff Juanita Valdez-Cox lives in Hidalgo County, Texas, where the 2010 census response rate was 55.9 percent. Valdez-Cox Decl. ¶ 48. When doorknocking follow-up began on August 9, 2020, Hidalgo County’s self-response rate was 48.5 percent, 7.4 points behind its final 2010 rate, and well behind the statewide rate. Valdez-Cox Decl. ¶ 51. As of August 30, Hidalgo County’s response rate was 51.4 percent, 3.5 percentage points behind its final 2010 rate and well behind the current 2020 national self-response rate of 64.9 percent. • Plaintiffs Debbie Chen and many members of OCA-GH live in Harris County, Texas, the most populous county in the nation’s second-most populous state (and the third-most populous county anywhere in the United States). When doorknocking follow-up began on August 9, 2020, Harris County’s self-response rate was 57.7 percent, 7.4 percentage points behind its final 2010 rate of 65.1 percent. As of August 30, Harris County’s response rate was 59.9 percent, 5.2 percent behind its final 2010 rate of 65.1 percent and well behind the current 2020 national self-response rate of 64.9 percent. • Plaintiff John Park is a resident of Queens County, New York. When doorknocking follow-up began on August 9, 2020, Queens County’s self-response rate was 55.1 percent, 5.7 percent behind its 2010 response rate. As of August 30, Queens County’s response rate was 58.8 percent, 2.0 percent behind its final 2010 rate of 60.8 percent and well behind the current 2020 national self-response rate of 64.9 percent. • Plaintiff Candy L. Gutierrez lives and votes in Yakima County, Washington. Declaration of Candy Gutierrez (“Gutierrez Decl.”), Ex. Q ¶¶ 2-6. Although the self-response rate of the county at large as of August 30 is 64.6 28 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 33 of 44 percent, slightly behind the national average and just over four percentage points behind the county’s 2010 response rate, the Yakama Nation tribe, comprising a substantial portion of the district’s population, has a 52.7 percent response rate, well behind its 2010 response rate of 58.1 percent and the current national rate of 64.9 percent. It is extremely likely that the Yakama Nation population will be severely undercounted in the census, and that non-White residents of Yakima County will therefore be undercounted as a group. • Plaintiff Petra Falcon, a member of Promise Arizona (PAZ) in Maricopa County, Arizona, lives and votes in South Phoenix, an area of Maricopa County. Although Maricopa County has a higher-than-average response rate as of this writing, Arizona’s Seventh Congressional District, a majority-Latino district encompassing South Phoenix, has a response rate of 54.1 percent as of August 30, 2020, trailing its 2010 response rate of 57.5 percent and substantially behind the national response rate of 64.9 percent. Plaintiffs have attested to the practical difficulties of data collection and fieldwork amid the COVID-19 pandemic, the diversity of languages spoken in Latino and Asian American communities and the challenges that such diversity presents to enumerators, and the adverse effects that the truncated data collection period will have upon census response outreach efforts. Gonzalez Decl. ¶¶ 9-12; Park Decl. ¶¶ 18-26; Chen Decl. ¶¶ 18-23; Valdez Decl. ¶¶ 10-17; Falcon Decl. ¶¶ 8-18; Valdez-Cox Decl. ¶¶ 18-23. In Queens County, New York, home to more than half of New York City’s large Asian American population, many populations are linguistically isolated and require contact in one of a variety of languages; many live in densely populated, hard-to-count areas; and many have developed a profound distrust of the government. Park Decl. ¶¶ 15-20. Community organizations, including Organizational Plaintiffs, are trusted in the community and have a proven track record of improving census responses. Park Decl. ¶ 10. The truncated timeline will undermine Organizational Plaintiffs’ outreach efforts, and make them less effective. The same is true of Harris County, Texas, and Maricopa County, Arizona, in which a substantial portion of the population resides in mobile homes or is homeless, further exacerbating the challenge of counting them. Chen Decl. ¶ 19; Falcon ¶¶ 7-10, 17-21. Rural areas face many of the same issues—linguistic isolation, distrust of government, lack of Internet 29 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 34 of 44 access—but also several unique ones; rural populations are less dense, requiring more extensive travel by enumerators, and many individuals also do not have a formal or permanent address at which they can receive mail, including the entire community of Las Lomas in the Rio Grande Valley. Valdez-Cox Decl. ¶¶ 7-8, 21-23, 27, 29-30. As Plaintiffs’ experts have testified, the differential in response rates can be substantially mitigated by a longer and more thorough follow-up process. Conversely, a truncated follow-up period will exacerbate the difference in response rates. Thompson Decl. ¶¶ 15-21; Hogan Decl. ¶¶ 8-17. If Defendants are permitted to move forward on the timeline they themselves so recently disavowed, the Census will ultimately yield an unprecedented undercount of the U.S. population. This undercount will have particularly harmful effects on Black, Latino, and Asian American communities. O’Hare Decl. ¶ 1 (“Changing the end date of data collection in the 2020 Census from October 31, 2020 to September 30, 2020, will impact the quality of Census data for Hispanic, Asians, and Blacks more than Non-Hispanic Whites.”). Specifically, many of the cities and counties in which Plaintiffs live are statistically certain to experience a significant undercount, with dire consequences. C. Precedent Demonstrates That Plaintiffs Are Entitled to Relief The Census is performed in obedience to a constitutional command. The Constitution requires that an “actual Enumeration ... be made within every subsequent Term of ten Years.” U.S. Const. art. I, § 2, cl. 3. Although the Constitution grants Congress sole authority over the census, Congress has delegated that power to the Secretary of Commerce to carry out the Census. The Secretary’s discretion is not, however, without limit. The Secretary’s actions in carrying out the Census must bear “a reasonable relationship to the accomplishment of an actual enumeration of the population.” Wisconsin v. City of New York, 517 U.S. 1, 20 (1996). The Secretary’s conduct must be “consistent with the constitutional language and the constitutional 30 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 35 of 44 goal of equal representation.” Id. at 19-20 (quoting Franklin, 505 U.S. at 804). “[W]hen the Census Bureau unreasonably compromises the distributive accuracy of the census, it may violate the Constitution.” La Unión del Pueblo Entero v. Ross, 353 F. Supp. 3d 381, 393 (D. Md. 2018). In Wisconsin, the Court analyzed whether the Secretary’s decision on the matter in dispute—whether to perform certain statistical adjustments—was consistent with “the constitutional purposes of the census, viz., to determine the apportionment of the Representatives among the States,” finding that the Secretary’s preference for distributive accuracy over numerical accuracy was consistent with that purpose. 517 U.S. at 20. The Court further emphasized “the importance of historical practice in this area,” finding that the Secretary’s decision was consistent with “the traditional method of conducting the census.” Id. at 22. Here, Defendants do not point to any changed circumstances that suggest truncation of enumeration is appropriate. Nothing suggests Defendants can complete an accurate count by the end of 2020. Census Bureau Director Steven Dillingham could offer no reason for the truncation when he testified to Congress other than to say that the Trump Administration wanted the Census concluded by the end of the year. In fact, Defendants’ own public statements demonstrate that the contemplated truncation of operations will prevent them from achieving “the constitutional purposes of the census,” id. at 20. Defendants have stated publicly and repeatedly that more time than the August 3 announcement provides is necessary to ensure “the completeness and accuracy of the 2020 Census.”54 Truncating field operations and data processing will prevent “accomplishment of an actual enumeration” through collection and processing of information from all households, with deleterious effects on Plaintiffs’ communities. See supra I.A-B. The 54 U.S. Census Bureau, Statement on 2020 Census Operational Adjustments Due to COVID-19 (Apr. 13, 2020), https://www.census.gov/newsroom/press-releases/2020/statementcovid-19-2020.html. 31 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 36 of 44 inaccurate counts that will result will significantly compromise the distributive accuracy of the Census and undermine the “constitutional goal of equal representation,” 517 U.S. at 19-20, by depriving Plaintiffs and other Latinos, Asian Americans and immigrants of equal representation. Nor can Defendants argue that their truncation plan has any foundation in “historical practice” or the “traditional method of conducting the census.” The global pandemic forced the extraordinary suspension of Census field operations. Defendants have now announced their intent to proceed with the shortest NRFU operation in modern history—52 days, compared to periods of three to six months in previous censuses.55 Defendants’ suggestions about how they might compensate for this departure from historical practice have been facially implausible and swiftly discredited. The August 3 announcement claimed the Census Bureau could—through measures including the “hiring of more employees”—“improve the speed of our count without sacrificing completeness,”56 but the Department’s Inspector General has warned that the Census Bureau has had difficulty “hiring and retaining sufficient numbers of field employees during the shortened nonresponse followup (NRFU) operation” and is accordingly “at risk of not conducting a complete and accurate 2020 Census.”57 Nor could increased staffing during postprocessing (even assuming a sufficient number of staff could be identified who possess the training needed to execute the required complex statistical analysis) compensate for hiring deficiencies at the NRFU stage or the compressed timeline. Hogan Decl. ¶ 17. Indeed, as one 55 Infra n.10. 56 See Statement from U.S. Census Bureau Director Steven Dillingham: Delivering a Complete and Accurate 2020 Census Count (Aug. 3, 2020), https://www.census.gov/newsroom/ press-releases/2020/delivering-complete-accurate-count.html. 57 U.S. Dep’t of Commerce, Office of Inspector General, 2020 Census Alert: The Census Bureau Faces Challenges in Accelerating Hiring and Minimizing Attrition Rates for Abbreviated 2020 Census Field Operations (Aug. 18, 2020), https://www.oig.doc.gov/OIGPublications/OIG20-041-M.pdf. See also Thompson Decl. ¶ 21. 32 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 37 of 44 expert has testified, the Bureau’s “policy setting and decision-making processes” are simply “unable to adapt quickly to address the effects of [COVID-19]” on Census operations—which have been “rattled” in multiple ways by the pandemic. O’Hara Decl. ¶ 8; id. ¶¶ 11-15. Defendants’ truncation of the timeframe to complete the Census is especially perverse in the midst of a global pandemic and at a time when natural disasters that have forced evacuations in many communities that will disrupt or prevent NRFU. The disruptions wrought by COVID19 are manifest. Meanwhile, more than 750,000 residents of Texas and Louisiana have been forced to evacuate their homes in the wake of Hurricane Laura.58 In California, more than 100,000 residents have been evacuated in the wake of widespread wildfires, and thousands of homes have been destroyed.59 And in one Florida city, Hurricane Michael caused approximately 30 percent of the population to relocate temporarily.60 These natural disasters are legitimate threats to an accurate census count. Brace Decl. ¶ 23. As its primary rationale for proceeding on such a truncated schedule, the August 3 announcement invoked the “statutory deadline of December 31, 2020.”61 This statutory deadline, see 13 U.S.C. § 141(b), cannot trump Defendants’ constitutional obligations. There is 58 Weather.com, Hurricane Laura: Two C-130s Land in Port Arthur, Texas, for “Last Chance” Evacuation; Louisiana Prepares for Storm Surge (Aug. 25, 2020), https://weather.com/news/news/2020-08-26-hurricane-laura-texas-louisiana-evacuationspreparations 59 Wyatte Grantham-Phillips and Joel Shannon, About 560 wildfires burn in California as more than 100,000 people evacuate: Here’s what we know (Aug. 21, 2020), https://www.usatoday.com/story/news/nation/2020/08/21/california-wildfires-how-manyburning-what-we-know/3406246001/. 60 Mark McQueen, The Hidden Post-Natural-Disaster Threat: the 2020 Census, Governing.com (May 19, 2020), https://www.governing.com/now/The-Hidden-Post-NaturalDisaster-Threat-the-2020-Census.html. 61 Statement from U.S. Census Bureau Director Steven Dillingham: Delivering a Complete and Accurate 2020 Census Count (Aug. 3, 2020), https://www.census.gov/newsroom/pressreleases/2020/delivering-complete-accurate-count.html. 33 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 38 of 44 “nothing sacred in the due date of the filing” of the total population figures with the President. Carey v. Klutznick, 637 F. 2d 834, 837 (2d Cir. 1980). This is especially so where, as here, “the work of the Census Bureau … is incomplete,” and where “a census undercount is inevitable,” “the undercount is particularly large among minority populations,” and “Census Bureau procedures were inadequate.” Id. at 837, 839. As here, the plaintiffs in Carey alleged the following irreparable harm arising out of an inevitable undercount of minority populations: “the dilution of the votes of [urban] residents, particularly members of minority groups vis-a-vis those of other residents of the state with respect to the state legislature, and … [loss of] vast sums of money distributed under federal revenue sharing and other programs with statutory formulas tied to the census.” Id. at 837. The evidence here is, if anything, stronger than that proffered in Carey. For example, although the plaintiffs in Carey successfully alleged that “the follow-up check … by the postal service and census workers was wholly inadequate,” the 1980 NRFU did not incorporate a historically short timeline with severe staffing shortages amid a global pandemic and a series of natural disasters. Congress has expressed no intent “to sacrifice accuracy for the sake of timeliness.” City of New York v. U.S. Dep’t of Commerce, 713 F. Supp. 48, 51 (E.D.N.Y. 1989). Thus, if the Census Bureau demonstrates that accurate adjusted figures cannot be compiled by December 31 … this Court is empowered to grant it a reasonable extension of time.” II. Truncation of Census Operations Will Cause Irreparable Harm to Plaintiffs As four former Census Bureau Directors have confirmed, the early cessation of census field operations and the associated limitations on data processing and quality check time will inevitably lead to an under-enumeration of those individuals who have not yet responded to the Census. That undercount will have a disparate impact on Black, Latino, and, in certain states, Asian American communities, which have, historically and to date (judging by self-response 34 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 39 of 44 rates in particular tracts with large populations of communities of color) demonstrated lower response rates than the nation as a whole. O’Hare Decl. ¶¶ 125-127. A. Loss of Political Representation Plaintiffs are at immediate and severe risk of losing political representation at the local, state, and national levels. Communities of color likely will be undercounted in the 2020 Census, and Defendants’ truncation of the timeline they themselves identified for completion will fatally exacerbate that undercount. The resulting differential undercount will, in turn, diminish the political representation of communities of color, including those in which Plaintiffs live, in state and local redistricting. Thompson Decl. ¶ 4; O’Hare Decl. ¶¶ 123-125; Brace Decl. ¶¶ 32-33. Furthermore, the undercount may well lead to malapportionment of congressional seats to some of the states in which plaintiffs live, such as Texas, Florida, and New York. Brace Decl. ¶¶ 2130; O’Hare Decl. ¶¶ 95-96. An undercount because of low self-response rates could cost a congressional seat and electoral vote in the State of Texas, to which Texas would otherwise be entitled. O’Hare Decl. ¶ 95; Brace Decl. ¶¶21-30. Diminished political representation in Congress is an immediate and irreparable injury of constitutional magnitude, if “there is a substantial risk that the harm will occur.” Dep’t of Commerce v. New York, 139 S. Ct. 2551, 2565, 204 L. Ed. 2d 978 (2019); Reynolds v. Sims, 377 U.S. 533, 560-63 (1964); Wesberry, 376 U.S. at 1, 7-8 (1964). Such is the case here, where the Defendants propose to end NRFU operations in just over one month, inevitably leaving communities of color undercounted and underrepresented in the 2021 apportionment. Thompson Decl. ¶ 22; see also Brace Decl. ¶¶32-33. But Plaintiffs’ representational injury does not stop at the congressional level. The anticipated undercount will have an independent effect of drawing Plaintiffs into artificially overpopulated intra-state voting districts, affecting representation at all levels. Brace Decl. ¶ 33. 35 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 40 of 44 States rely on census data to draw state legislative districts, and census undercounts result in seats shifting away from regions in which people are undercounted.62 O’Hare Decl. ¶ 99; Brace Decl. ¶ 33. For example, as a result of relatively faster population growth, Cameron and Hidalgo counties, located in the Rio Grande Valley at the southern tip of Texas, gained a seat in the state House of Representatives following the 2010 Census.63 As of August 30, 2020, self-response rates in Cameron and Hidalgo counties are 49.3 percent and 51.4 percent respectively—well behind the Texas self-response rate of 60.3 percent.64 The Rio Grande Valley is also currently in the grip of the COVID-19 pandemic, with case rates higher than the national average as well as other areas of Texas. Currently, Cameron County has witnessed 4,694 cases per 100,000 residents, while Hidalgo County has seen 2,890 cases per 100,000 residents.65 The significant undercount that will inevitably result from truncating enumeration in Cameron and Hidalgo counties may well result in the loss of the recently-gained state house seat, or the loss of an additional seat in 2021 to which the Rio Grande Valley would otherwise be entitled, if the undercount eliminates the population that contributed to the additional seat. The truncation of field operations and data processing operations will therefore diminish Plaintiffs’ political representation at every level of government, from local to national elections. 62 See, e.g., Ariz. Const. art. 4, Pt. 2, § 1 (3); Cal. Const. art. 21, § 1; Fla. Stat. § 11.031(1); N.J. Const. art. 4, § 2 ¶¶ 1, 3; Tex. Const. art. 3, § 26; Kravitz v. United States Dep’t of Commerce, 366 F. Supp. 3d 681, 737 (D. Md. 2019) (where “a differential undercount of [a particular population] of some magnitude is virtually certain to occur” because of a challenged action, that action “creates the substantial risk that certain Plaintiffs will suffer vote dilution.”). 63 Perez v. Texas, 891 F. Supp. 2d 808, 813 (W.D. Tex. 2012). 64 U.S. Census Bureau, Response Rates (Aug. 25, 2020), https://2020census.gov/en/response-rates.html. 65 The New York Times, Coronavirus in the U.S.: Latest Map and Case Count (Aug. 25, 2020), https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html#states. 36 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 41 of 44 B. Loss of Public Resources The harmful effects of an undercount are neither abstract nor speculative; they are real, severe, and imminent. One effect of an undercount concerns eligibility for and allocation of federal assistance, often targeted toward the same underserved communities in which many Plaintiffs live. As Plaintiff’s expert has set forth, more than three hundred distinct federal programs rely on census data to allocate over $1.5 trillion of domestic assistance funding annually. Thompson Decl. ¶¶ 3, 13; Declaration of Andrew Reamer (“Reamer Decl.”), Ex. F, ¶ 14. Some examples of federal programs that rely on census data for funding allocations include Medicaid, Medicare, the State Children’s Health Insurance Program (CHIP), Social Security block grants, the Supplemental Nutrition Assistance Program (SNAP), the National School Lunch Program, and the Supplemental Income Program for Women, Infants, and Children (WIC). Many in Plaintiffs’ communities, including Organizational Plaintiffs’ members, rely on social services funded by such programs, including without limitation Medicaid, food stamps or SNAP, CHIP, Community Development Block Grants, Section 8 housing, and Title I schools. Falcon Decl. ¶ 21; Valdez-Cox Decl. ¶¶ 10-11. Those programs use census data to determine eligibility requirements (i.e., who may receive federal funding); geographic allocations, or which areas of the country are most in need of federal funding; selection criteria (for example, federal project applications may be scored on the basis of population density or diversity); and interest rates for federal loan programs (which can be determined, for example, on the basis of median household income in a particular area). Reamer Decl. ¶¶ 15-20. As Plaintiffs’ expert has explained, states, counties, and communities that are undercounted will suffer a loss of federal funding to which they are otherwise entitled. Reamer Decl. ¶¶ 12-13. If Defendants’ current timeline remains in place, the economic impact in terms of federal funding will be devastating—particularly in view of the fact that the programs identified above 37 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 42 of 44 are specifically targeted toward underserved communities, including children, the disabled, the elderly, and the economically disadvantaged. Those are often precisely the same communities that are undercounted in the Census. Falcon Decl. ¶ 21; Valdez-Cox Decl. ¶¶ 10-11. Moreover, as set forth above, several of these communities have been among the hardest hit by COVID19and are reliant on federally funded programs using census data, such as Medicaid, Medicare, and CHIP, for medical assistance. See, e.g., Valdez-Cox Decl. ¶¶ 10-11. For example, New York City has seen more than 238,000 diagnoses and over 23,000 deaths,66 and diagnoses are well above state and national averages in the Rio Grande Valley of Texas. See supra at 35-36; Valdez-Cox Decl. ¶¶ 40-46. A differential undercount will prevent these communities from receiving the assistance they critically need in the wake of the pandemic, both with respect to medical assistance programs generally and COVID-specific programs with allocations based on census data. Invisibility in census data therefore harms those who need federal assistance the most. III. The Balance of Equities and the Public Interest Support a Temporary Restraining Order or Preliminary Injunction When the government is opposing a TRO or preliminary injunction, the analyses of the balance of equities and the public interest merge. J.O.P. v. U.S. Dep’t of Homeland Sec., 409 F. Supp. 3d 367, 376 (D. Md. 2019) (citing Nken v. Holder, 556 U.S. 418 (2009)). Here, Plaintiffs have shown numerous personal harms, including “loss of federal funding and vote dilution,” and have further shown a broader public interest in preventing Defendants from “undermin[ing] the Census’s primary constitutional purpose by unreasonably distorting the accuracy of ‘an actual Enumeration.’” Kravitz v. United States Dep’t of Commerce, 366 F. Supp. 3d 681, 755 (D. Md. 66 The New York Times, New York Coronavirus Map and Case Count (last updated Aug. 30, 2020), https://www.nytimes.com/interactive/2020/us/new-york-coronavirus-cases.html. 38 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 43 of 44 2019). Defendants can offer no plausible rationale supporting the decision to truncate enumeration and the truncation itself is undermined by “the Census Bureau’s own data and analyses.” Id. at 751-52. In such circumstances, this Court has held, the balance of equities and the public interest weigh in favor of injunctive relief. CONCLUSION Plaintiffs respectfully request that the Court enter a temporary restraining order enjoining Defendants from halting the self-response period and census field collection operations prior to October 31, 2020, and reporting the count of total population to the President prior to April 30, 2021. 39 Case 8:19-cv-02710-PX-PAH-ELH Document 112-1 Filed 09/01/20 Page 44 of 44 Dated: September 1, 2020 Respectfully submitted, By /s/ Terry Ao Minnis ASIAN AMERICANS ADVANCING JUSTICE AAJC John C. Yang (IL Bar No. 6210478)* Niyati Shah (NJ Bar No. 026622005) Terry Ao Minnis (MD Bar No. 20547)º Eri Andriola (NY Bar No. 5510805)º 1620 L Street, NW, Suite 1050 Washington, DC 20036 Phone: (202) 815-1098 Facsimile: (202) 296-2318 MEXICAN AMERICAN LEGAL DEFENSE AND EDUCATIONAL FUND Thomas A. Saenz (CA Bar No. 159430)º Nina Perales (TX Bar No. 24005046)º Andrea Senteno (NY Bar. No. 5285341)º Tanya G. Pellegrini (CA Bar No. 285186)º Daniel A. Hatoum (TX Bar No. 24099136)º 1016 16th Street NW, Suite 100 Washington, DC 20036 Phone: (202) 293-2828 Facsimile: (202) 293-2849 WILMER CUTLER PICKERING HALE AND DORR LLP Debo P. Adegbile (NY Bar No. 2699742) Jamie S. Dycus (NY Bar No. 4523569) Alan E. Schoenfeld (NY Bar. No. 4500898) 7 World Trade Center 250 Greenwich Street New York, NY 10007 Phone: (212) 230-880 Facsimile: (212) 230-8888 * Pro hac vice application forthcoming º Not admitted in DC. 40