Case No. 20-CV-05799-LHK
ORDER DENYING DEFENDANTS’ MOTION TO DISMISS AND MOTION FOR STAY OF PROCEEDINGS
Re: Dkt. Nos. 354, 355
LUCY H. KOH, United States District Judge
Plaintiffs National Urban League; League of Women Voters; Black Alliance for Just Immigration; Harris County, Texas; King County, Washington; City of Los Angeles, California; City of Salinas, California; City of San Jose, California; Rodney Ellis; Adrian Garcia; National Association for the Advancement of Colored People; City of Chicago, Illinois; County of Los Angeles, California; Navajo Nation; and Gila River Indian Community (collectively, "Plaintiffs") sue Defendants Commerce Secretary Wilbur L. Ross, Jr.; the U.S. Department of Commerce; the Director of the U.S. Census Bureau Steven Dillingham, and the U.S. Census Bureau ("Bureau") (collectively, "Defendants") for violations of the Enumeration Clause and the Administrative Procedure Act ("APA").
Before the Court are Defendants’ motion to dismiss Plaintiffs’ second amended complaint, ECF No. 354, and Defendants’ motion for stay of proceedings, ECF No. 355. Having considered the parties’ submissions; the relevant law; and the record in this case, the Court DENIES Defendants’ motion to dismiss and motion for stay of proceedings.
A. Factual Background
The 2020 Census is "a 15.6 billion dollar operation years in the making." ECF No. 81. As a result, after nearly a decade of preparation, Defendants adopted a final operational plan for the 2020 Census in December 2018 called the Operational Plan Version 4.0. However, in March 2020, shortly after the beginning of data collection, the COVID-19 pandemic upended Defendants’ Operational Plan and necessitated more time for census operations. Accordingly, on April 13, 2020, Defendants adopted the COVID-19 Plan, which elongated the schedule for data collection and processing and the Secretary of Commerce's reports of population "tabulations" to the President and the states. See 13 U.S.C. § 141(b), (c). On August 3, 2020, Defendants announced the Replan, which reduced the COVID-19 timeframes for data collection and processing by half. Below, the Court first describes census data collection, data processing, and reporting in general terms. The Court then details the deadlines for these operations under the Operational Plan Version 4.0; the COVID-19 Plan; and the Replan. Lastly, the Court recounts reports on the Replan by the Government Accountability Office; the Bureau's Scientific Advisory Committee; and the Department of Commerce's Office of Inspector General.
1. Deadlines for data collection, data processing, and the Secretary's reports to the President and the states.
As relevant here, there are four key deadlines in the 2020 Census. First is the deadline for self-responses to census questionnaires. At the end of the self-response period, the Census Bureau stops accepting responses to the census.
Second is the deadline on which Non-Response Follow-Up ("NRFU") ceases. NRFU refers to the process of "conduct[ing] in-person contact attempts at each and every housing unit that did not self-respond to the decennial census questionnaire." Fontenot Decl. ¶ 48, ECF No. 81-1. "The NRFU Operation is entirely about hard-to-count populations." ECF No. 37-5 at 219. NRFU is thus "the most important census operation to ensuring a fair and accurate count." Thompson Decl. ¶ 15, ECF No. 36-2. Together, self-responses and NRFU comprise the census's data collection.
Third is the deadline for data processing after data collection. Data processing refers to the Bureau's "procedures to summarize the individual and household data that [the Bureau] collect[s] into usable, high quality tabulated data products." Fontenot Decl. ¶ 66.
Lastly, at the end of data collection and processing, the Secretary of Commerce issues two reports pursuant to the Census Act: (1) "the tabulation of total population by States" for congressional apportionment to the President by December 31, 2020, see 13 U.S.C. § 141(b) ; and (2) a tabulation of population for redistricting to the states by April 1, 2021, see id. § 141(c).
2. The Operational Plan Version 4.0, adopted in December 2018, provided a total of 54 weeks for the 2020 Census.
Defendants’ declarant, Albert E. Fontenot, Jr., Associate Director for Decennial Census Programs at the U.S. Census Bureau, describes the Bureau's extensive work over nearly a decade to develop the Operational Plan Version 4.0 (hereafter, "Operational Plan"). For example, Associate Director Fontenot discusses eight significant census tests the Bureau performed in 2013, 2014, 2015, 2016, and 2018 to improve their field operations. Fontenot Decl. ¶ 71. Associate Director Fontenot describes partnerships with stakeholders such as organizations, tribes, and local governments. E.g. , Fontenot Decl. ¶¶ 12, 28. The Operational Plan reflects the conclusions of subject-matter experts such as statisticians, demographers, geographers, and linguists. See, e.g. , ECF No. 37-5 at 79, 144 (2020 Census Operational Plan—Version 4.0).
For an organizational chart of the Census Bureau, see Census Bureau Organizational Chart, https://www.census.gov/about/who.html, ECF No. 150-3. Director Steven Dillingham and Deputy Director Ron Jarmin head the Bureau, and their direct reports are Associate Directors.
Under the Operational Plan adopted in December 2018, self-responses spanned 20.5 weeks from March 12 to July 31, 2020. NRFU spanned 11.5 weeks from May 13 to July 31, 2020. Data processing spanned 22 weeks from August 1 to December 31, 2020. These operational dates would culminate in the Secretary of Commerce issuing his reports by the statutory deadlines. Specifically, by December 31, 2020, the Secretary would report "the tabulation of total population by States" to the President for the purpose of Congressional apportionment. By April 31, 2021, the Secretary would report the tabulation of population to the states for the purpose of redistricting. 13 U.S.C. § 141(b).
3. COVID-19 pandemic causes suspension of census operations.
Six days after the self-response period began on March 12, 2020, the Bureau announced on March 18, 2020 that it would suspend all field operations for two weeks because of the COVID-19 pandemic. See Press Release, U.S. Census Bureau, U.S. Census Bureau Director Steven Dillingham on Operational Updates (Mar. 18, 2020), https://www.census.gov/newsroom/press-releases/2020/operational-update.html.
The Bureau foresaw an eight-week operational delay, according to an internal Bureau document dated March 24, 2020 and sent by the Bureau Deputy Director's Chief Advisor, Enrique Lamas, to senior staff. The document stressed the importance of maintaining an uncompressed schedule. Reasons for maintaining an uncompressed schedule included completing the workload remaining and operations that ensured a complete count of all population groups:
• The document stated that "staff had covered only about 10% of the workload when [the Bureau] had to stop." DOC_7087.
• The document further noted that operations "focused on counting populations not living in traditional housing, such as nursing home residents, college students, the military, prisoners, the homeless, and the transitory populations are being planned and will be conducted as it is safe for Census employees and the public to engage in face-to-face activities. These operations and our nonresponse follow-up operation, all need to be completed before the Census Bureau can begin processing the data to ensure that we have a complete count of the population and not undercount specific population groups." DOC_7088.
In line with the Bureau's expectation of a long delay, the Bureau announced another two-week suspension on March 28, 2020. Press Release, Census Bureau Update on 2020 Census Field Operations (Mar. 28, 2020), https://www.census.gov/newsroom/press-releases/2020/update-on-2020-census-field-operations.html. Further delays followed.
Ultimately, the Bureau's projected eight-week delay was nine weeks plus phased restarts. The Chief of Staff to Secretary Ross, Michael Walsh, analyzed the issues for the Secretary on May 8, 2020. Walsh wrote that "[p]ursuant to OMB guidance, the Census Bureau completely suspended decennial field operations for 47 days between March 18 and May 4," and then resumed operations in phases thereafter. DOC_2287 (emphasis in original) ("Operational Timeline" memo). Walsh flagged issues with two operations especially important to avoiding undercounts, enumerator onboarding and "Update Leave":
• Onboarding enumerators "entails recruitment, selection, acceptance and gathering of any additional information, fingerprinting, background checks, onboarding, and training" approximately 340,000–500,000 enumerators. Id. "The suspension of field operations curtailed preparation for
this [onboarding], as much of it required personal contact." Id. After onboarding, enumerators "visit non-responding households and conduct in-person interview to obtain census responses." DOC_2287.
• Update Leave, as Walsh wrote, "helps reach 5 million homes in the USA in rural and remote areas that lack city-style mail." Id. Update Leave reaches those homes by having Census "field staff hand-deliver questionnaires," id. at 6, to "areas where the majority of the housing units do not have mail delivery ... or the mail delivery information for the housing unit cannot be verified." Fontenot Decl. ¶ 46. Before the complete suspension of operations, "approximately 10% of the initial [Update Leave] workload had been completed." DOC_2287. By contrast, "[u]nder initial projections, 100% of the Update Leave workload should have been completed by April 17." Id.
The May 8, 2020 Operational Timeline memo also foresaw problems with "[d]ata processing and integrity." Id. (emphasis omitted). "[T]he pandemic has made impacts that will require additional processing and expertise because populations have temporarily shifted." Id. As a result, the memo suggested that the 2018 Operational Plan's provision of 152 days (about 22 weeks) for data processing was not enough. Id.
As field operations began restarting under the COVID-19 Plan detailed below, the Bureau encountered COVID-related challenges. In particular, the Bureau had trouble retaining enumerators and conducting in-person visits in NRFU. On retaining enumerators, Associate Director for Field Operations Timothy Olson wrote to other senior officials on July 23, 2020 that "[the Bureau] had a huge quit rate from training to deployed in field (and this does not mirror past censuses at all – it is MUCH higher, almost a debilitating higher quit rate). And this translate[d] into much slower production in the field because we have less than half the number of enumerators (38%) we need to get the job done." DOC_7737.
Issues with NRFU visits were flagged in a June 10, 2020 presentation sent by the Chief of Staff to Director Dillingham, Christa Jones, to Deputy Director Jarmin and the Chief of Staff to the Deputy Secretary of Commerce, Dan Risko. DOC_6545. On a slide titled "Risks and Challenges Due to COVID-19," the presentation stated that COVID-19 had "le[ ]d to new risks and unknowns for the operation." Id. Four risks stood out: (1) a lower case resolution rate because respondents "may be less likely to answer their door"; (2) challenges with staffing and training; (3) a complex schedule; (4) and a "de-scoped" early NRFU operation that presumably had been delayed by COVID. Id.
By July 30, 2020—by which time the Bureau had already been directed to create the Replan, as discussed below—enumerator staffing was still low. DOC_8623. Many cities across several Area Census Offices had roughly 50% shortfalls in enumerator staffing compared to the Bureau's internal target. Id. Plaintiffs’ affidavits allude to similar issues with finding enumerators. In Monterey County, California, for instance, the pandemic made it harder to hire and retain enumerators "because traditional applicant groups like senior citizens have concerns about the risk of catching COVID-19." Gurmilan Decl. ¶ 13.
4. The COVID-19 Plan, adopted on April 13, 2020, provided 71.5 weeks for the 2020 Census.
As a result, on April 13, 2020, the Bureau issued an adjustment to its Operational Plan to account for the impact of COVID-19 (the "COVID-19 Plan"). ECF No. 37-3 (April 13, 2020 statement of Secretary of Commerce Wilbur Ross and Census Bureau Director Steven Dillingham). The COVID-19 Plan extended the deadlines. Specifically, first, the COVID-19 Plan expanded the deadlines for self-responses from 20.5 weeks to 33.5 weeks (March 12 to October 31, 2020) to account for the pandemic's disruptions to Bureau operations and the public's ability to respond to the census. Second, NRFU likewise expanded from 11.5 weeks (May 13 to July 31, 2020) to 12 weeks (August 11 to October 31, 2020).
Third, given the pandemic's effects on "the quality of the data, especially for groups that are less likely to self-respond (often hard to count populations)," post-data collection quality control was deemed especially important. ECF No. 37-7 at 18. Data processing for congressional apportionment thus expanded from 22 weeks (August 1 to December 31, 2020) to 26 weeks (November 1, 2020 to April 30, 2021). The processing was to include an independent review of the final address list, analysis by subject-matter experts, and the remediation of software errors. Fontenot Decl. ¶ 89.
Lastly, the press release announcing the COVID-19 Plan stated that "the Census Bureau is seeking statutory relief from Congress of 120 additional calendar days to deliver final apportionment counts." ECF No. 37-3 at 3. The COVID-19 Plan would thus "extend the window for field data collection and self-response to October 31, 2020, which will allow for apportionment counts to be delivered to the President by April 30, 2021, and redistricting data to be delivered to the states no later than July 31, 2021." Id.
Although these delays would result in the Bureau missing statutory deadlines, the President of the United States and Bureau officials publicly stated that meeting the December 31, 2020 deadline would be impossible in any event. On the day the COVID-19 Plan was announced, President Donald J. Trump stated, "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. They have to give it. I think 120 days isn't nearly enough." ECF No. 131-16 at 4.
On May 14, 2020, the Bureau's Chief of Intergovernmental Affairs, Chris Stanley, sent a memo to Secretary Ross on what Ross would say to mayors about the 2020 Census. DOC_14997. According to the memo, Secretary Ross would call mayors on May 15, 2020 to tell them that "there is still more than enough time for your residents to self-respond. Due to the coronavirus, we have extended the time for planned census field operations for the 2020 Census by 3 months – from July 31 to October 31." Id.
On May 26, 2020, the Bureau's Associate Director for Field Operations, Timothy Olson, stated that "[w]e have passed the point where we could even meet the current legislative requirement of December 31. We can't do that anymore. We -- we've passed that for quite a while now." Nat'l Conf. of Am. Indians, 2020 Census Webinar: American Indian/Alaska Native at 1:17:30–1:18:30, YouTube (May 26, 2020), https://www.youtube.com/watch?v=F6IyJMtDDgY.
Likewise, on July 8, Associate Director Fontenot confirmed that the Bureau is "past the window of being able to get" accurate counts to the President by December 31, 2020. U.S. Census Bureau, Operational Press Briefing – 2020 Census Update at 20–21 (July 8, 2020), https://www.census.gov/content/dam/Census/newsroom/press-kits/2020/news-briefing-program-transcript-july8.pdf. The Bureau's internal view on missing the statutory deadlines was similar. Days after announcing the COVID-19 Plan, the Bureau prepared for a call on April 28, 2020 with Congressman Jamie Raskin, Chair of the House Oversight Subcommittee on Civil Rights and Civil Liberties, which has jurisdiction over the census. In preparation for that call, the Bureau's Chief of Congressional Affairs, Christopher Stanley, circulated a memo to Director Dillingham and other senior officials. See DOC_2224. The memo answered possible questions about missed deadlines.
Two questions and answers ("Q&As") stood out. The first Q&A contemplated that any data collection after August 14 would make meeting the deadlines infeasible. The Q&A asked why the Bureau couldn't "collect data after August 14 and still deliver redistricting data on time?" DOC_2227. The answer was that the Bureau had "examined [the] schedule and compressed it as much as [the Bureau] c[ould] without risking significant impacts on data quality. Given the important uses of census data collection processing, it is vital that [the Bureau] not shortcut these efforts or quality assurance steps." Id.
The second Q&A asked whether "delaying the apportionment data [was] constitutional?" The answer was that "[t]he proposal underwent a constitutional review, and we believe it is constitutional and that the adjusted schedule will help us fulfill the constitutional requirement of a complete and accurate census.... In history, especially for the many of the earlier censuses, data collection and reporting the counts shifted beyond the zero year." DOC_2228. By "counts shifted beyond the zero year," the Bureau presumably was referring to census reports that had been made in the calendar year after the statutory deadline. Those reports were for the censuses of 1810, 1820, 1830, and 1840. ECF No. 203 (explaining examples); see, e.g. , Act of Sept. 1, 1841, ch. 15, § 1, 5 Stat. 452, 452 (second post hoc extension of September 1, 1841 for original deadline missed by over nine months). In those censuses, after one or more deadlines had passed without the enumeration having been completed, Congress extended the relevant deadlines after the fact. See ECF No. 203.
Moreover, without a deadline extension from Congress, the Census Bureau has twice missed the statutory deadline for completing data collection, but continued collecting data. Specifically, in the 1940 and 1950 Censuses, Congress provided only two weeks for the enumeration of large cities (i.e. , cities with population over 250,000) and 30 days for the rest of the country. See Pub. L. No. 71-13, § 6, 46 Stat. 21, 23 (1929) ("Provided , that in any city having [2500,000] inhabitants or more under the preceding census[,] the enumeration of the population shall be completed within two weeks from the commencement thereof."). Even so, the Bureau continued data collection because meeting those deadlines would have led to an inaccurate count. See Census Bureau, The 1950 Censuses—How They Were Taken at 19 (1955), https://www2.census.gov/prod2/decennial/documents/1950/proceduralHistory/00322994ch04.pdf ("In a few districts, this schedule had to be modified because of bad weather, floods, recruitment problems, short working hours, and unsatisfactory enumeration."); The United Press, 17,000,000 Still Unlisted As Census Taking Lags , N.Y. Times (Apr. 24, 1940), https://timesmachine-nytimes-com.sw.library.ntpu.edu.tw:8443/timesmachine/1940/04/24/92949988.html?pageNumber=1 ("Census officials estimated today that they still had 17,000,000 persons to count.... Several weeks’ more work will be required to find every one."). Unlike in the 1810 through 1840 Censuses, Congress never extended the deadlines for the 1940 and 1950 Census after the fact. On May 8, 2020, Secretary Ross's Chief of Staff, Michael Walsh, sent the "Operational Timeline" memo to the Secretary. The Operational Timeline memo found that:
If the Census Bureau could fully restart today, under ideal conditions ... the earliest you could finish NRFU, even with the ability to restart immediately every state, is approximately September 1, 2020. By finishing NRFU on September 1, 2020, apportionment counts could not be delivered until January 31, 2021, already after the statutory deadline. Redistricting information would be provided to states by April 30, 2021, already after the statutory deadline.
Based on the initial suspension of field activities in line with OMB guidance, the Census Bureau can no longer meet its statutory deadlines for delivering apportionment and redistricting data, even conducting operations under unrealistically ideal conditions.
DOC_2288 (emphasis in original) (bullet points omitted).
All the above operational concerns were ultimately reflected in the census response data. As of June 2020, "self-response rates var[ied] widely across states and counties," with "markedly different operational environments and challenges" facing the Bureau "from one locale to another." ECF No. 37-7 at 6 (citing self-response rates "below 3 percent" in counties in Alaska, Texas, Utah, and South Dakota).
5. The Replan, adopted on August 3, 2020, reduced the time for the 2020 Census from 71.5 weeks to 49.5 weeks.
On July 20, 2020, the Bureau's Chief Financial Officer, Ben Page, asked other senior officials whether the Bureau still supported Congressional extension of the statutory deadlines. DOC_6852 (July 20, 2020 email to Director Dillingham et al.). Page wrote:
Among the first questions I am getting is "Does the Census bureau still need the change in the statutory dates?" Can we find a time to discuss how we should respond to that question? Given that the Senate may introduce a bill today or tomorrow, I anticipate we'll need a set answer for discourse over the next 24-48 hours.
On July 21, 2020, President Trump issued a memorandum declaring the United States’ policy to exclude undocumented immigrants from the congressional apportionment base. See Excluding Illegal Aliens From the Apportionment Base Following the 2020 Census , 85 Fed. Reg. 44,679 (July 21, 2020) ("Presidential Memorandum").
On July 23, 2020, Associate Director Fontenot started an email thread with several senior Bureau officials, including Deputy Director Ron Jarmin and Associate Director for Field Operations Timothy Olson. Associate Director Fontenot began the thread by stating that on July 27, he would tell the Department of Commerce about the "reality of the COVID Impacts and challenges":
On Monday at DOC [Department of Commerce] I plan to talk about the difference between goal and actual case enumeration (Currently a shortfall (11 % goal vs 7% actual) and attribute it to the higher drop out rate and (ideally with reasons) and what we are going to do to address the technology drop outs.)
I think it is critical to lay the groundwork for the reality of the COVID Impacts and challenges.
Does anyone have any problems with my approach?
DOC_7737. In response, Associate Director Olson "agree[d] that elevating the reality is critical, especially in light of the push to complete NRFU asap for all the reasons we know about." DOC_7738. Those reasons are not in the administrative record.
Associate Director Olson then "sound[ed] the alarm" on "deliver[ing] apportionment by 12/31" in the strongest possible terms:
We need to sound the alarm to realities on the ground – people are afraid to work for us and it is reflected in the number of enumerators working in the 1a ACOs [Area Census Offices]. And this means it is ludicrous to think we can complete 100% of the nation's data collection earlier than 10/31 and any thinking person who would believe we can deliver apportionment by 12/31 has either a mental deficiency or a political motivation.
Id. One reason that accelerating the schedule would be "ludicrous," Associate Director Olson stated, was the "awful deploy rate" of enumerators about 62% below target. Id. Driving that shortfall was "almost a debilitating higher quit rate":
Another tack is to provide crystal clear numbers by the 1a ACOs that shows the awful deploy rate - field selected the right number (big number) to training, training show rate was on par with prior censuses (albeit a few points lower ... but overall in line with past censuses). And then we had a huge quit rate from training to deployed in field (and this does not mirror past censuses at all - it is MUCH higher, almost a debilitating higher quit rate). And this translates into much slower production in the field because we have less than half the number of enumerators (38%) we need to get the job done.
On the same day as Associate Director Olson's email (July 23, 2020), the Chief of Decennial Communications and Stakeholder Relationships, Kathleen Styles, shared a so-called "Elevator Speech" memo with Government Accountability Office ("GAO") official Ty Mitchell and senior Bureau officials. See DOC_8026 (email to GAO). The purpose of the Elevator Speech, Chief Styles wrote, was "to explain, in layman's terms, why we need a schedule extension." The Speech begins with a "High Level Message," which in its entirety reads:
Curtailing census operations will result in a census that is of unacceptable quality. The Census Bureau needs the full 120 days that the Administration originally requested from Congress to have the best chance to produce high quality, usable census results in this difficult time. Shortening the time period to meet the original statutory deadlines for apportionment and redistricting data will result in a census that has fatal data quality flaws that are unacceptable for a Constitutionally-mandated activity.
Still, on July 29, 2020 at a House Oversight and Reform hearing, Director Dillingham did not support extending the statutory deadline. Rather, he sidestepped questions about whether the "Administration has [ ] reversed direction on [the extension], and is now suggesting that they want the Census to be wrapped up quickly so that th[e] tabulation ... could actually happen before the end of the year." Oversight Committee, Counting Every Person at 3:50:42–3:51:40, YouTube (July 29, 2020), https://youtu.be/SKXS8e1Ew7c?t=13880 (questions by Congressman John Sarbanes). Director Dillingham's response was that "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." Id. at 3:51:48–3:52:02. On July 30, 2020 at 9:18 a.m., Associate Director Fontenot wrote Associate Director Olson, Olson's deputy (Assistant Director for Field Operations James T. Christy), and other Bureau leadership with an update on "condensing the back end of the schedule to hit the Statutory dates for the Apportionment count." DOC_15654. Associate Director Fontenot asked Olson and Christy if field operations could end by September 30, 2020. Specifically, Associate Director Fontenot wrote: "[s]o that the plan comes together would hope that you can also work toward a 9/30 end date to be out of the field.... If you don't mind, let's use the Operations meeting this morning to discuss the first two objectives on the plan – Accelerating the start of Field operations and maintaining staff levels." Id. At 9:44 a.m., Assistant Director Christy responded: "[a]n impossible task, but we will try." Id. Seven minutes later, Associate Director Olson also chimed in: "I'm with Jamey on this. We talked at length last night." DOC_15653.
Also on July 30, 2020, Deputy Director Jarmin wrote to Bureau leadership with instructions from "KDK"—Deputy Secretary of Commerce Karen Dunn Kelley. Deputy Director Jarmin stated that "[a]s we prepare stuff for tomorrow morning KDK has stressed we need to address" four points, including "[m]eeting requirements of the PM (and EO re CVAP)." DOC_15718. The "PM" is the July 21, 2020 Presidential Memorandum on excluding undocumented immigrants from the congressional apportionment base. "EO re CVAP" is Executive Order 13,880, which directs the government to develop data on the citizen voting-age population. See Collecting Information About Citizenship Status in Connection With the Decennial Census , Fed. Reg. 33,821, 33823–25 (July 11, 2019).
On the evening of July 30 and the twilight hours of July 31, 2020, senior Bureau officials debated how to respond to NPR's questions about why "the Census Bureau has moved up the end of 2020 census door knocking from Oct. 31 to Sept. 30." DOC_15736. In response to NPR's question whether "the decision was made in order to meet the current statutory deadline of Dec. 31," Associate Director Fontenot suggested answering "[w]e have not received relief from our statutory requirements." DOC_15734. Yet Associate Director for Communications Ali Ahmad disagreed with that proposed answer. As Associate Director Ahmad explained, "Al, while true that we haven't received the relief[,] we are also backing away from asking for it. So I think that answer doesn't quite work in that fashion, because the follow up doesn't have a great response." Id.
On July 31, 2020, the Bureau removed from its website the October 31, 2020 deadline for data collection without any announcement or explanation. Compare ECF No. 37-8 (July 30 Operational Adjustments Timeline), with ECF No. 37-9 (July 31 Operational Adjustments Timeline).
By August 1, 2020, the Bureau had prepared several versions for a presentation to Secretary Ross on Monday, August 3, 2020 ("August 3 Presentation"). The parties identify one version as a key document. ECF Nos. 161 at 2 (Defendants’ identification of DOC_10275), 190 at 6 (Plaintiffs’ identification of same). The Presentation's very first slide, titled "Overview," concludes that "to achieve an acceptable level of accuracy, at least 99% of Housing Units in every state must be resolved":
Due to COVID-19 impacts, the conclusion of field operations for the 2020 Census was previously scheduled to end on October 31. In order to meet the statutory
date of December 31, 2020 for apportionment, field operations must now conclude no later than September 30, 2020. Accelerating the schedule by 30 days introduces significant risk to the accuracy of the census data. In order to achieve an acceptable level of accuracy, at least 99% of Housing Units in every state must be resolved.
On August 3, 2020, the Bureau issued a press release announcing a "new plan," which the Bureau called the "Replan." U.S. Census Bureau, Statement from U.S. Census Bureau Director Steven Dillingham: Delivering a Complete and Accurate 2020 Census Count (Aug. 3, 2020), ECF No. 37-1 ("August 3 Press Release"). In his declaration, Associate Director Fontenot avers that the Secretary approved the Replan on the day it was announced. Fontenot Decl. ¶ 85.
In the words of the August 3 Press release, the Replan "accelerate[d] the completion of data collection and apportionment counts by our statutory deadline of December 31, 2020, as required by law and directed by the Secretary of Commerce." ECF No. 37-1. The time for the 2020 Census was reduced from 71.5 weeks to 49.5 weeks. Specifically, self-response compressed from 33.5 weeks to 29 weeks, with the deadline advancing from October 31 to September 30. Fontenot Decl. ¶ 100. NRFU compressed from 11.5 weeks to 7.5 weeks, with the deadline advancing from October 31 to September 30. Lastly, data processing was halved from 26 weeks to 13 weeks, with the deadline advancing from April 30, 2021 to December 31, 2020.
As of August 3, 2020, less than 63% of households had responded to the 2020 Census. ECF No. 37-1.
6. The Government Accountability Office found that the Replan increases the risks to obtaining a complete and accurate 2020 Census.
In June 2020, the GAO issued a Report on the 2020 Census entitled, "COVID-19 Presents Delays and Risks to Census Count," in which the GAO noted, among other things, that staffing shortages were experienced at the Bureau's call centers and at the Bureau's contractor responsible for printing the six mail-in self-response forms. ECF No. 37-7 at 8 (GAO, COVID-19 Presents Delays and Risks to Census Count (June 2020)). The Report also noted that as of June 1, 2020, counties in Alaska, Texas, Utah, and South Dakota had reported self-response rates below 3 percent. Id. at 9.
The Court may take judicial notice of matters that are either "generally known within the trial court's territorial jurisdiction" or "can be accurately and readily determined from sources whose accuracy cannot reasonably be questioned." Fed. R. Evid. 201(b). Courts take judicial notice of information, such as reports of the Government Accountability Office ("GAO"), Census Scientific Advisory Committee ("CSAC"), and Department of Commerce Office of Inspector General ("OIG"), which are found on government agency websites. See Paralyzed Veterans of Am. v. McPherson , 2008 WL 4183981, at *5–6 (N.D. Cal. Sept. 9, 2008) (citing circuit and district court cases). However, to the extent any facts in the documents subject to judicial notice are subject to reasonable dispute, the Court will not take judicial notice of those facts. See Lee v. City of L.A. , 250 F.3d 668, 689 (9th Cir. 2001) ("A court may take judicial notice of matters of public record.... But a court may not take judicial notice of a fact that is subject to reasonable dispute.") (internal quotation marks omitted), overruled on other grounds by Galbraith v. Cty. of Santa Clara , 307 F.3d 1119 (9th Cir. 2002).
In August 2020, the GAO issued a Report on the 2020 Census entitled "Recent Decision to Compress Census Timeframes Poses Additional Risks to an Accurate Count." https://www.gao.gov/assets/710/709015.pdf. The Report stated: "Delays to data collection operations, public reluctance to participate in door-to-door interviews, and compressed timeframes for data collection and processing response data may affect the accuracy, completeness, and quality of the count." Id. at ii (cover memo). The Report also noted that implementation of untested procedures and continuing challenges such as COVID-19 could "undermine the overall quality of the count." Id. at 1.
The reports of the GAO, CSAC, and OIG are not in the administrative record. In adjudicating Plaintiffs’ APA claims, the Court is permitted to go outside the administrative record "for the limited purpose of background information." Thompson v. U.S. Dep't of Labor , 885 F.2d 551, 555 (9th Cir. 1989). The Court does not consider the reports for APA analysis. That said, many of the documents on which the OIG Report is based are included in the partial administrative record, which is the basis of the Court's APA analysis. Moreover, the Court's analysis of Plaintiffs’ Enumeration Clause claim is not limited to the administrative record. Thus, for the Enumeration Clause claim, the Court may consider these reports for more than just background information.
7. The Bureau's Scientific Advisory Committee unanimously supports extension of the census schedule.
Associate Director Fontenot's September 22, 2020 declaration states: "In the midst of major West Coast fires and air quality issues that have accelerated since September 11, and the current impacts of Hurricane Sally across the states of Louisiana, Mississippi, Alabama, the Florida panhandle area, parts of Georgia, and South Carolina, I stated publicly on September 17, 2020 in the Census Scientific Advisory Committee meeting that I did not know whether Mother Nature would allow us to meet the September 30 date." ECF No. 196-1 ¶ 14.
The next day, on September 18, 2020, the Census Scientific Advisory Committee ("CSAC") unanimously concluded that the Census schedule should be extended. See Allison Plyer, Census Scientific Advisory Committee Chair, Recommendations and Comments to the Census Bureau from the Census Scientific Advisory Committee Fall 2020 Meeting (September 18, 2020), https://www.documentcloud.org/documents/7213520-Recommendations-and-Comments-to-the-Census.html#document/p2/a581794. Specifically, the CSAC found the following:
To ensure a successful completion of the 2020 Census in a way that is consistent with its mandate of counting everyone once and in the right place, and based on its scientific and methodological expertise, CSAC recommends that the 2020 Census operational timeline be extended per the Bureau's April 2020 request. Counting everyone once and in the right place, using untested and never-before-used technologies, that must work together with precision, requires time. When the weather isn't right, we postpone the launching of rockets into space. The same should be true of the decennial enumeration, the results of which will impact apportionment, redistricting, funding decisions, legal mandates and regulatory uses of decennial Census data over the next decade.
Id. at 2.
8. The Commerce Department's Office of Inspector General found that the Replan increases the risks to obtaining a complete and accurate 2020 Census.
On September 21, 2020, the Department of Commerce's Office of Inspector General ("OIG") released a report entitled "The Acceleration of the Census Schedule Increases the Risks to a Complete and Accurate 2020 Census." Final Management Alert No. OIG-20-050-M (Sept. 18, 2020), https://www.oig.doc.gov/OIGPublications/OIG-20-050-M.pdf. The Report drew upon Bureau and Commerce Department documents that were produced to the OIG (the "OIG production" stated below), as well as interviews with senior Bureau officials and Director Steven Dillingham. Id. at 2. The report made two findings. First, "[t]he decision to accelerate the Census schedule was not made by the Census Bureau." Information Memorandum for Secretary Ross from Peggy E. Gustafson at 1 (Sept. 18, 2020). Second, "[t]he accelerated schedule increases the risks to obtaining a complete and accurate 2020 Census." Id.
On the first finding, the report detailed that:
As of mid-July 2020, the Bureau still viewed the statutory extension as necessary in order to conduct the 2020 Census completely and accurately. This view is consistent with previous public statements made by senior Bureau officials that the Bureau would no longer be able to meet the December 31, 2020, statutory deadline.
Then, in the late afternoon of Wednesday, July 29, 2020, a senior Department official told the Bureau to put together options for meeting the apportionment deadline of December 31, 2020, and brief the Secretary on those options on Monday morning, August 3, 2020.
Id. at 7. On the second finding, the report detailed that "senior Bureau officials believed that the largest risk to data collection posed by the accelerated plan was the decreased time to recover from possible external contingencies affecting local areas or regions." Id. at 8.
As of September 21, 2020, the Census Bureau had resolved 99% of housing units in only four states. ECF No. 196-1 ¶ 13. The Bureau had stated internally in its August 3 Presentation that "[i]n order to achieve an acceptable level of accuracy, at least 99% of Housing Units in every state must be resolved." DOC_1026. Thus, with only nine days left before the close of data collection under the Replan, the Bureau was missing its own requirements for state completion rates for 46 states.
The Court notes these later extra-record developments for context, but does not weigh them in its APA analysis. But cf. Dep't of Commerce v. New York , ––– U.S. ––––, 139 S. Ct. 2551, 2575, 204 L.Ed.2d 978 (2019) ("It is rare to review a record as extensive as the one before us when evaluating informal agency action—and it should be.... [B]ut we are ‘not required to exhibit a naiveté from which ordinary citizens are free.’ " (quoting United States v. Stanchich , 550 F.2d 1294, 1300 (2d Cir. 1977) (Friendly, J.))).
B. Procedural History
On September 24, 2020, after extraordinarily expedited proceedings recounted at length in other orders, the Court stayed and preliminarily enjoined the Replan under the APA. ECF No. 208 ("Injunction Order"). Because Plaintiffs’ APA arbitrary-and-capricious claim was dispositive in five independent ways, the Court did not need to reach Plaintiffs’ Enumeration Clause claim or APA pretext claim. See Injunction Order at 44, 46.
Much has transpired since the Injunction Order issued. As described below, Defendants violated the Injunction Order. Then, Defendants sought a stay of the Injunction Order from the Ninth Circuit and ultimately the Supreme Court. Now, Defendants have filed the instant motions and began discovery. The Court addresses each procedural stage in turn.
1. Defendants violate the Court's injunction.
On September 25, 2020, Defendants appealed to the Ninth Circuit. ECF No. 210. There, Defendants sought (1) an immediate administrative stay of the Injunction Order pending resolution of Defendants’ stay motion; (2) a stay pending appeal; (3) and review on the merits. See Nat'l Urban League v. Ross , No. 20-16868 (9th Cir.) (case docket).
Meanwhile, Defendants repeatedly violated the Injunction Order. ECF No. 288 (detailing violations). Perhaps the most egregious violation of the Injunction Order occurred on Monday, September 28, 2020. At 1:58 p.m., two minutes before the Court's case management conference, the Census Bureau tweeted one sentence: "The Secretary of Commerce has announced a target date of October 5, 2020 to conclude 2020 Census self-response and field data collection operations." @USCensusBureau, https://twitter.com/uscensusbureau/status/1310685274104569856. Later, the Census Bureau issued a one sentence press release with the exact same sentence. See U.S. Census Bureau, 2020 Census Update (Sept. 28, 2020), https://www.census.gov/newsroom/press-releases/2020/2020-census-update.html. Neither the one sentence tweet nor the one sentence press release provided any explanation or information. The Court thus ordered Defendants to produce the administrative record of this announcement. ECF No. 225.
The production showed three significant things. First, Defendants set the October 5 date to meet the December 31, 2020 statutory deadline, even though Defendants were "enjoined from implementing" that deadline. Census Bureau Deputy Director Ron Jarmin presented to Secretary Ross two "Proposed Options for Completion of Enumeration"—both of which focused on the December 31, 2020 deadline that the Court had stayed and enjoined Defendants from implementing:
Option 1 : Conclude field work by October 5, 2020 in order to meet apportionment delivery date of December 31, 2020.
Option 2 : Continue field work beyond October 5, 2020 in order to increase state completion rates to 99% and to continue to improve enumeration of lagging sub-state areas, such as tribal areas, rural areas, and hard-to-count communities. However, this would not allow for delivery of state counts for apportionment by December 31, 2020.
ECF No. 233 at 148 (italics added). As Deputy Director Jarmin explained to Director Dillingham and other senior officials, Option 2 "would preclude meeting the 12/31 date, but furthers the goal of a complete and accurate 2020 Census. " Id. at 130 (emphasis added). Option 1, by contrast, would not further that goal.
To pick between the two options (ending data collection by or after October 5, 2020), Secretary Ross asked which would implement the December 31, 2020 deadline. Specifically, on September 28, 2020 at 3:52 p.m. Eastern, Secretary Ross wrote to Deputy Director Jarmin and other senior Bureau officials: "As I prepare to make the decision, I would like to make sure that I understood correctly that your team's opinion is that if we stay in the field beyond October 5, we would not be able to meet the statutory deadline of December 31." ECF No. 256-1 at 2. At 4:30 p.m. Eastern, Deputy Director Jarmin responded: "Yes sir, we need to finish field work on 10/5 if we are to have enough time (and assuming all goes well) to finish the processing of the resident population, federally affiliated overseas and, if requested, unlawful aliens in ICE Detention Centers by 12/31. Other PM [Presidential Memorandum] related outputs would be pushed to 1/11/2021." Id. at 1. ECF No. 256-1. Second, Deputy Director Jarmin's response to Secretary Ross made clear that the December 31, 2020 statutory deadline intertwined with the President's July 21, 2020 Memorandum on Excluding Illegal Aliens from the Apportionment Base Following the 2020 Census.
Third, Secretary Ross approved the October 5 date 14 minutes after the Census Bureau tweeted the October 5 date. At 5:12 p.m. Eastern—14 minutes after the Bureau's tweet announcing the Secretary's decision—Secretary Ross wrote back: "Thanks for the confirmation. Based on the staff recommendation I am extending the field operation to October [sic] 5." Id.
Thus, on October 1, 2020, the Court (1) clarified the Injunction Order; (2) ordered certain limited relief to preserve the status quo; and (3) warned Defendants that the Court would subject them to sanctions or contempt proceedings if Defendants violated the Injunction Order again. ECF No. 288 at 14–15.
On September 30, 2020—the day data collection would have ended but for the Injunction Order—the reported completion rate of 16 states and the District of Columbia was still under 99%. See U.S. Census Bureau, 2020 Census Housing Unit Enumeration Progress by State (Oct. 1, 2020 report date), https://2020census.gov/content/dam/2020census/news/daily-nrfu-rates/nrfu-rates-report-10-01.pdf.
2. The Ninth Circuit rejects all but one of Defendants’ arguments for a stay; and the Supreme Court stays the Injunction Order on October 13, 2020.
Also on September 30, 2020, the Ninth Circuit denied Defendants’ application for an immediate administrative stay. See Nat'l Urban League v. Ross ("Nat'l Urban League I "), 977 F.3d 698, 703 (9th Cir. 2020). The Ninth Circuit held that issuing an administrative stay would "disband[ ] thousands of census workers," thereby "eliminating the Bureau's ability to conduct field work." Id. at 701. Thus, "the status quo would be upended, rather than preserved, if an administrative stay is issued" Id. at 703. United States Circuit Judge Patrick J. Bumatay dissented. Id. Judge Bumatay argued that the panel "should have granted an administrative stay here because defendants are likely to succeed on the merits." Id. at 705 (Bumatay, J., dissenting).
On October 7, 2020, a different Ninth Circuit panel unanimously denied Defendants’ motion for a stay in part and granted it in part. Nat'l Urban League v. Ross ("Nat'l Urban League II "), 977 F.3d 770, 773 (9th Cir. 2020). The Ninth Circuit left in place the injunction against "the Replan and the September 30, 2020[ ] deadline for data collection." Id. The Ninth Circuit rejected Defendants’ arguments on APA reviewability and the merits. On reviewability, Defendants failed to show that they were likely to prevail on their argument that the Replan was neither "agency action" nor "final." See id. at 776–77 (discussing both parts of "final agency action"). On APA merits, the Ninth Circuit held that "[t]he district court laid out in great detail five grounds on which to find Plaintiffs were likely to succeed on their contention that the government did not meet the APA's standards for reasoned decisionmaking." Id. at 777 ; see id. at 777–79 (discussing grounds).
However, "[t]o the extent that the district court enjoined the government from attempting to meet the December 31, 2020, statutory deadline for completing tabulations by state," the Ninth Circuit stayed the injunction. Id. The Ninth Circuit reasoned, among other things, that "predictions as to whether [the December 31, 2020 deadline] can still be attained are speculative and unstable. And any harm from governmental attempts to meet the December 31 date are likely less irreparable than the injury from displacing the October 31 data collection endpoint." Id. at 781.
Defendants then sought a stay from the Supreme Court. See Ross v. National Urban League , ––– U.S. ––––, 141 S.Ct. 18, 208 L.Ed.2d 169 (2020) (case docket), https://www.supremecourt.gov/docket. Defendants represented to the Supreme Court that, with a stay, Defendants could still meet the December 31, 2020 deadline with a 99% completion rate. See Defs. Stay App. at 8, National Urban League (U.S. Oct. 7, 2020) (citing "sworn testimony in the record ... of the government's ability to meet the statutory deadline on its proposed schedule"); id. at 13 (discussing completion rate and irreparable harm). On October 13, 2020, the Supreme Court stayed the Injunction Order "pending disposition of the appeal in the United States Court of Appeals for the Ninth Circuit and disposition of the petition for a writ of certiorari." National Urban League , ––– U.S. ––––, ––––, 141 S.Ct. 18, 208 L.Ed.2d 169 (2020) (slip op. at 1). The Supreme Court's stay order did not provide reasons for granting the stay. Defendants have since represented to the Supreme Court that they can no longer meet the December 31, 2020 deadline. See Oral Arg. Tr. at 6:16–7:3, Trump v. New York , No. 20-366 (U.S. Nov. 30, 2020) ("[W]e are not currently on pace to send the report to the President by the year-end statutory deadline.").
3. Defendants file the instant motions to dismiss and to stay, and the parties start discovery.
On October 27, 2020, Plaintiffs filed the Second Amended Complaint. ECF No. 352 ("SAC"). On November 10, 2020, Defendants filed the instant motions (1) to dismiss the SAC; and (2) to stay proceedings. ECF Nos. 354 ("motion to dismiss" or "mot."), 355 (motion to stay).
On November 12, 2020, the parties filed a joint case management statement. ECF No. 356. The joint case management statement proposed competing case schedules. Plaintiffs sought discovery on their Enumeration Clause claim, which this Court had not addressed in prior orders. Plaintiffs asked this Court for an expedited schedule with (1) a one-month discovery period; and (2) final judgment before the Secretary's report of congressional apportionment figures to the president, which Plaintiffs expected sometime in early January 2021 (after the December 31, 2020 statutory deadline). ECF No. 356 at 9. By contrast, Defendants proposed that (1) the case be stayed until the Bureau completed the 2020 Census or until Defendants’ appeal of the Court's preliminary injunction were resolved; and (2) following a stay, there be a three-month discovery period. Id. at 10.
Starting with the Court's November 13, 2020 Case Management Order and two following orders, the Court explained why discovery should proceed and why such discovery should be limited. ECF Nos. 357 (Case Management Order), 372 (Dec. 10, 2020 Order to Compel), 380 (Dec. 13, 2020 Order Denying Motion for Reconsideration). The Court set an expedited schedule because of the states’ "contingent redistricting deadlines" and the difficulty of granting relief after the April 1, 2020 deadline for the Secretary's transmittal of redistricting data to the states. This schedule gave the parties eight weeks to engage in limited discovery. ECF No. 357 at 2. Because of the expedited schedule, the Court gave the parties 10 days to respond to Requests for Production and 14 days to start producing documents for which no objections were asserted. Id. The Court also significantly limited discovery in several ways. For instance, although the Federal Rules of Civil Procedure (the "Federal Rules") permit 10 depositions per side, the Court cut that number nearly in half, permitting only 6 depositions. See Fed. R. Civ. P. 30(a)(1) ; ECF Nos. 357 (case management order), 372 (Order to Compel). Similarly, the Federal Rules permit 25 interrogatories per side, but the Court cut that number by 60%, permitting only 10 interrogatories. See Fed. R. Civ. P. 33(a)(1) ; ECF No. 357. Lastly, the Federal Rules permit an unlimited number of Requests for Admission ("RFAs") and Requests for Production ("RFPs"), but the Court permitted only 25 RFAs and 25 RFPs per side in the instant case. See Fed. R. Civ. P. 34, 36 ; ECF No. 357.
On November 24, 2020, Plaintiffs filed their oppositions to Defendants’ motion to dismiss and motion to stay. ECF Nos. 364 (opposition to motion to dismiss, or "Opp'n"), 365 (opposition to motion to stay). Defendants did not file a reply in support of either motion.
II. LEGAL STANDARD
A. Motion to Dismiss Under Rule 12(b)(1)
A motion to dismiss pursuant to Federal Rule of Civil Procedure 12(b)(1) tests whether the court has subject matter jurisdiction. Although lack of "statutory standing" requires dismissal for failure to state a claim under Rule 12(b)(6), lack of Article III standing requires dismissal for want of subject matter jurisdiction under Rule 12(b)(1). See Nw. Requirements Utilities v. F.E.R.C. , 798 F.3d 796, 808 (9th Cir. 2015) ("Unlike Article III standing, however, ‘statutory standing’ does not implicate our subject-matter jurisdiction." (citing Lexmark Int'l, Inc. v. Static Control Components, Inc. , 572 U.S. 118, 128 n.4, 134 S.Ct. 1377, 188 L.Ed.2d 392 (2014) )); Maya v. Centex Corp. , 658 F.3d 1060, 1067 (9th Cir. 2011). A Rule 12(b)(1) jurisdictional attack may be factual or facial. Safe Air for Everyone v. Meyer , 373 F.3d 1035, 1039 (9th Cir. 2004).
"[I]n a factual attack, the challenger disputes the truth of the allegations that, by themselves, would otherwise invoke federal jurisdiction." Id. In resolving such an attack, unlike with a motion to dismiss under Rule 12(b)(6), a court "may review evidence beyond the complaint without converting the motion to dismiss into a motion for summary judgment." Id. Moreover, the court "need not presume the truthfulness of the plaintiff's allegations." Id. Once the defendant has moved to dismiss for lack of subject matter jurisdiction under Rule 12(b)(1), the plaintiff bears the burden of establishing the court's jurisdiction. See Chandler v. State Farm Mut. Auto Ins. Co. , 598 F.3d 1115, 1122 (9th Cir. 2010).
"In a facial attack," on the other hand, "the challenger asserts that the allegations contained in a complaint are insufficient on their face to invoke federal jurisdiction." Safe Air for Everyone , 373 F.3d at 1039. The court "resolves a facial attack as it would a motion to dismiss under Rule 12(b)(6) : Accepting the plaintiff's allegations as true and drawing all reasonable inferences in the plaintiff's favor, the court determines whether the allegations are sufficient as a legal matter to invoke the court's jurisdiction." Leite v. Crane Co. , 749 F.3d 1117, 1121 (9th Cir. 2014).
B. Leave to Amend
If a court determines that a complaint should be dismissed, it must then decide whether to grant leave to amend. Under Rule 15(a) of the Federal Rules of Civil Procedure, leave to amend "shall be freely given when justice so requires," bearing in mind "the underlying purpose of Rule 15 to facilitate decisions on the merits, rather than on the pleadings or technicalities." Lopez v. Smith , 203 F.3d 1122, 1127 (9th Cir. 2000) (en banc) (alterations and internal quotation marks omitted). When dismissing a complaint for failure to state a claim, "a district court should grant leave to amend even if no request to amend the pleading was made, unless it determines that the pleading could not possibly be cured by the allegation of other facts." Id. at 1130 (internal quotation marks omitted).
Accordingly, leave to amend generally shall be denied only if allowing amendment would unduly prejudice the opposing party, cause undue delay, or be futile, or if the moving party has acted in bad faith. Leadsinger, Inc. v. BMG Music Publ'g , 512 F.3d 522, 532 (9th Cir. 2008). At the same time, a court is justified in denying leave to amend when a plaintiff "repeated[ly] fail[s] to cure deficiencies by amendments previously allowed." See Carvalho v. Equifax Info. Servs., LLC , 629 F.3d 876, 892 (9th Cir. 2010). Indeed, a "district court's discretion to deny leave to amend is particularly broad where plaintiff has previously amended the complaint." Cafasso, U.S. ex rel. v. Gen. Dynamics C4 Sys., Inc. , 637 F.3d 1047, 1058 (9th Cir. 2011) (quotation marks omitted).
C. Motion to Stay Under Landis
Under Landis v. North American Co. , 299 U.S. 248, 57 S.Ct. 163, 81 L.Ed. 153 (1936), the Court has "discretionary power to stay proceedings in its own court." Lockyer v. Mirant Corp. , 398 F.3d 1098, 1109 (9th Cir. 2005). "A trial court may, with propriety, find it is efficient for its own docket and the fairest course for the parties to enter a stay of an action before it, pending resolution of independent proceedings which bear upon the case." Leyva v. Certified Grocers of California, Ltd. , 593 F.2d 857, 863 (9th Cir. 1979). "This rule applies whether the separate proceedings are judicial, administrative, or arbitral in character, and does not require that the issues in such proceedings are necessarily controlling of the action before the court." Id. "Where it is proposed that a pending proceeding be stayed, the competing interests which will be affected by the granting or refusal to grant a stay must be weighed." CMAX, Inc. v. Hall, 300 F.2d 265, 268 (9th Cir. 1962) (citing Landis , 299 U.S. at 254–55, 57 S.Ct. 163 ). As such, when determining whether to grant a stay, the Court weighs three factors: (1) the possible damage to the non-moving party, (2) "the hardship or inequity which a party may suffer in being required to go forward," and (3) "the orderly course of justice." CMAX , 300 F.2d at 268 ; see also Lockyer , 398 F.3d at 1110 (applying Landis factors).
The Court first addresses Defendants’ motion to dismiss. Defendants move to dismiss on five grounds for lack of jurisdiction under Federal Rule of Civil Procedure 12(b)(1). The first three grounds apply to all the claims in the SAC. Specifically, Defendants argue that (1) the political question doctrine bars Plaintiffs’ claims; (2) Plaintiffs lack Article III standing; and (3) Plaintiffs’ claims are not ripe. The last two grounds, which apply to Plaintiffs’ APA claims only, are that (4) the Replan is not final agency action; and (5) the Census Bureau's "choices" (which are not "agency action") are committed to agency discretion. Mot. at i.
Except for the ripeness argument, Defendants’ motion to dismiss raises the same threshold arguments that have already been rejected by this Court and the Ninth Circuit. This Court has repeatedly ruled in Plaintiffs’ favor on Defendants’ threshold issues. See, e.g. , Order to Produce the Administrative Record, ECF No. 86, at 9–17 (rejecting Defendants’ threshold arguments); Injunction Order, ECF No. 208, at 21–44 (same). Similarly, the Ninth Circuit has twice reviewed this Court's Injunction Order. The first Ninth Circuit panel denied Defendants’ motion for an administrative stay of the Injunction Order. Nat'l Urban League I , 977 F.3d at 703. The second Ninth Circuit panel expressly rejected Defendant's APA threshold arguments and implicitly rejected Defendants’ other threshold arguments. See Nat'l Urban League II , 977 F.3d at 773 (rejecting the "government's primary argument," which is lack of final agency action).
Below, the Court addresses all five arguments of Defendants’ motion to dismiss in turn. For ease of reference, the Court restates much of its analysis from the Injunction Order. After resolving the motion to dismiss, the Court then addresses Defendants’ motion for a stay.
A. This lawsuit still does not present a political question.
Defendants again argue that this lawsuit is not justiciable because it presents a political question. Mot. at 5–10. The Court still disagrees.
A "political question" is one which is "outside the courts’ competence and therefore beyond the courts’ jurisdiction." Rucho v. Common Cause , ––– U.S. ––––, 139 S. Ct. 2484, 2494, 204 L.Ed.2d 931 (2019). Tellingly, Defendants fail to offer a case that finds that the political question doctrine bars review of decisions regarding the administration of the census. Instead, Defendants point the Court to two defining hallmarks of a political question: " a textually demonstrable constitutional commitment of the issue to a coordinate political department; or  a lack of judicially discoverable and manageable standards for resolving" the dispute. Baker v. Carr , 369 U.S. 186, 217, 82 S.Ct. 691, 7 L.Ed.2d 663 (1962) ; accord Vieth v. Jubelirer , 541 U.S. 267, 277–78, 124 S.Ct. 1769, 158 L.Ed.2d 546 (2004). Defendants argue that both are present here because (1) the Enumeration Clause vests Congress with the authority to conduct "actual Enumeration," Mot. at 6–7, and (2) there is no evident standard by which the Court could evaluate the Bureau's decision. Mot. at 8–10. Neither argument is convincing.
First, Defendants still cite no case—and the Court is aware of none—in which a court declined jurisdiction over a census case on political question grounds. To the contrary, the Supreme Court and lower courts have repeatedly rejected the argument that the political question doctrine bars review of census-related decisionmaking. See, e.g. , U.S. Dep't of Commerce v. Montana , 503 U.S. 442, 458–59, 112 S.Ct. 1415, 118 L.Ed.2d 87 (1992) (holding that the "political question doctrine presents no bar"); Franklin v. Massachusetts , 505 U.S. 788, 801 n.2, 112 S.Ct. 2767, 120 L.Ed.2d 636 (1992) (noting that the Court "recently rejected a similar argument" in Montana that "the courts have no subject-matter jurisdiction over this case because it involves a ‘political question’ "); Carey v. Klutznick , 637 F.2d 834, 838 (2d Cir. 1980) (per curiam) (rejecting the Census Bureau's argument that "allegations as to mismanagement of the census made in the complaint involve a political question," and holding the case reviewable under the Constitution and APA) (quotation omitted); New York v. U.S. Dep't of Commerce , 315 F. Supp. 3d 766, 791 (S.D.N.Y. 2018) (rejecting political question doctrine in citizenship question litigation; and collecting cases); Young v. Klutznick , 497 F. Supp. 1318, 1326 (E.D. Mich. 1980) (rejecting political question doctrine), rev'd on other grounds , 652 F.2d 617 (6th Cir. 1981) ; City of Philadelphia v. Klutznick , 503 F. Supp. 663, 674 (E.D. Pa. 1980) (same); Texas v. Mosbacher , 783 F. Supp. 308, 312 (S.D. Tex. 1992) (same); District of Columbia v. U.S. Dep't of Commerce , 789 F. Supp. 1179, 1185 (D.D.C. 1992) (same); City of N.Y. v. U.S. Dep't of Commerce , 739 F. Supp. 761, 764 (E.D.N.Y. 1990) (same); U.S. House of Representatives v. U.S. Dep't of Commerce , 11 F. Supp. 2d 76, 95 (D.D.C. 1998) (three-judge court) (same; and stating "the court sees no reason to withdraw from litigation concerning the census"), aff'd , 525 U.S. 316, 119 S.Ct. 765, 142 L.Ed.2d 797 (1999) ; see also Utah v. Evans , 536 U.S. 452, 122 S.Ct. 2191, 153 L.Ed.2d 453 (2002) (engaging in review without noting any jurisdictional defect stemming from political question doctrine); Wisconsin v. City of N.Y. , 517 U.S. 1, 116 S.Ct. 1091, 134 L.Ed.2d 167 (1996) (same); Morales v. Daley , 116 F. Supp. 2d 801 (S.D. Tex. 2000) (same), aff'd sub nom. Morales v. Evans , 275 F.3d 45 (5th Cir. 2001) (unpublished); Prieto v. Stans , 321 F. Supp. 420, 421 (N.D. Cal. 1970) (finding jurisdiction over a motion to preliminarily enjoin the census's "mail-out, mail-back procedure" and "community education and follow-up procedures").
Second, precedent supports the determination that there is a discoverable and manageable standard by which the Court can review the agency action at issue here. For example, the Census Act "imposes ‘a duty to conduct a census that is accurate and that fairly accounts for the crucial representational rights that depend on the census and the apportionment.’ " Dep't of Commerce v. New York , ––– U.S. ––––, 139 S. Ct. 2551, 2569, 204 L.Ed.2d 978 (2019) (quoting Franklin , 505 U.S. at 819–820, 112 S.Ct. 2767 (Stevens, J., concurring in part and concurring in judgment)) (discussing 2 U.S.C. § 2a ). Similarly, the text, structure, and history of the Constitution evinces "a strong constitutional interest in accuracy." Utah , 536 U.S. at 455–56, 122 S.Ct. 2191.
Thus, in its decision on the census citizenship question last year, the Supreme Court rejected Defendants’ claim that there is "no meaningful standard against which to judge the agency's exercise of discretion." Dep't of Commerce v. New York , 139 S. Ct. at 2568 (quoting Weyerhaeuser Co. v. United States Fish and Wildlife Serv. , ––– U.S. ––––, 139 S. Ct. 361, 370, 202 L.Ed.2d 269 (2018) ). The standard is provided by the Census Act, the Constitution, and APA. Accordingly, it is no surprise that Defendants do not cite, and the Court could not find, a case in which the political question doctrine barred judicial review of census-related decisionmaking.
In sum, the political question doctrine does not bar the Court from reviewing the instant case.
B. Plaintiffs still have standing to challenge the Replan.
Next, Defendants again argue that Plaintiffs lack standing. Defendants were incorrect before and are incorrect now.
"To have standing, a plaintiff must ‘present an injury that is concrete, particularized, and actual or imminent; fairly traceable to the defendant's challenged behavior; and likely to be redressed by a favorable ruling.’ " Dep't of Commerce v. New York , 139 S. Ct. at 2565. Plaintiffs here allege—and support with affidavits—the same four injuries that the Supreme Court found supported standing in the citizenship question case: "diminishment of political representation, loss of federal funds, degradation of census data, and diversion of resources." Id. at 2565 (upholding findings as not clearly erroneous).
The Court first discusses each of Plaintiffs’ four alleged injuries. Then, the Court explains why Defendants’ supposed 99% completion rate fails to deprive Plaintiffs of standing. Lastly, the Court explains why Plaintiffs’ injuries are still traceable and redressable.
1. Plaintiffs are likely to lose federal funds that turn on census data.
The administrative record shows that the Replan will likely lead to an undercount that results in "loss of crucial federal funds for programs that affect [Plaintiffs’] daily life." A. Garcia Decl. ¶ 4. The Supreme Court has specifically agreed that the loss of federal funding "is a sufficiently concrete and imminent injury to satisfy Article III." Dep't of Commerce v. New York , 139 S. Ct. at 2565. Thus, the Court agrees that the possible loss of federal funds is a sufficient injury to establish Article III standing as explained below.
Local government Plaintiffs are recipients of multiple sources of federal funding that turn on census data. King County, Washington; the City of Los Angeles; and Harris County, Texas are leading examples. The Replan's shortened schedule for data collection and processing will likely diminish each locality's funding because each locality has many hard to count persons who risk being undercounted. M. Garcia Decl. ¶¶ 7–8; Dively Decl. ¶ 5; Briggs Decl. ¶¶ 7, 11; see also Hillygus Decl. ¶¶ 12, 19, 39 (explaining the statistics of undercounting subpopulations). Specifically, the Court notes the following:
• In King County, three-quarters of the County's record population growth of 15% since 2010 is attributable to "populations that are less likely to self-respond to the census." Dively Decl. ¶ 5. As a result, "[s]hortening the enumeration period risks creating a population undercount." Id. Any undercount would reduce King County's allocation of funds "proportionately disbursed by census population counts." Id. ¶ 7. These funds include Community Development Block Grants, HOME Investment Partnership Program, and Emergency Solutions Grants from the U.S. Department of Housing and Urban Development. Id. ¶ 7. Transit Formula Grants to the Seattle region, of which King County is a part, also turn on census data, and totaled $108 million in fiscal year 2019.
• Los Angeles County is "the hardest to count in the nation." M. Garcia Decl. ¶ 7. 57% of the residents in the City of Los Angeles, which is home to roughly 4 million people, live in census block groups that are hard or very hard to count. Id. As a result, Los Angeles’ self-response rate of 54.5% (as of August 19, 2020) is well below the city's 2010 response rate of 68% and the state's 2020 response rate of 65.9%.
• "[T]he City of Los Angeles receives tens of millions of dollars from the federal government each year based upon the ratio of population derived from the decennial census." Westall Decl. ¶ 35. In times of national emergency, cities such as Los Angeles receive relief based on census population. Id. ¶ 34 (discussing $20 million received under the Coronavirus Aid, Relief, and Economic Security Act, or CARES Act).
• In Harris County, the Replan's shortening of the self-response and NRFU timelines risks causing "unprecedented undercounts in the 2020 Census." Briggs Decl. ¶ 11. "[A]pproximately $90,529,359 of the grants expended by Harris County in
FY2019 depended on accurate census data." Wilden Decl. ¶ 5. Among the grants affected are those that enable "sustainable financing of local health departments" such as Harris County Public Health, which has helped manage COVID-19 for approximately 4.7 million people. Shah Decl. ¶¶ 4, 8.
An undercount in any locality matters greatly. Even a small undercount of a subset of the hard to count population would result in the loss of federal funding. See Dep't of Commerce v. New York , 139 S. Ct. at 2565 ("[I]f noncitizen households are undercounted by as little as 2% ... [states] will lose out on federal funds"). Thus, like in Department of Commerce v. New York , Plaintiffs that receive federal funds based on census population suffer "a sufficiently concrete and imminent injury to satisfy Article III." Id.
2. Plaintiffs will likely be deprived of their fair share of political representation.
Plaintiffs allege that the undercount resulting from the Replan will likely result in an unfair apportionment that will deprive local government Plaintiffs, individual Plaintiffs, and members of organizational Plaintiffs of their fair share of representation. The resulting "threat of vote dilution," whether Congressional or intrastate, is an injury in fact. Dep't of Commerce v. U.S. House of Representatives , 525 U.S. 316, 331–32, 119 S.Ct. 765, 142 L.Ed.2d 797 (1999).
For example, given the historically low census response rates in the City of Los Angeles and City of Salinas in California, the Replan creates a substantial risk that their residents will not be counted, and a substantial risk of diminished political representation. See M. Garcia Decl. ¶¶ 8–15; Gurmilan Decl. ¶¶ 6, 8–14. Specifically:
• In the City of Los Angeles, the Replan "will result in extreme inaccuracy" because it would leave "just over six weeks to complete enumeration of roughly half of the exceptionally diverse households of the nation's second-most-populous city—in the midst of a once-in-a-lifetime pandemic." M. Garcia Decl. ¶ 8; see Westall Decl. ¶ 36 (stating it is "likely" that undercounts will "disproportionally impact Los Angeles" and "cause the City to miss out on a portion of [ ] funding for an entire decade").
• Similarly, the City of Salinas comprises 38.5% of Monterey County's hard to count population, and the City's response rate is 9.5% below its response rate from the 2010 Census and 8% below the current state average. Gurmilan Decl. ¶ 6.
The undercount wrought by the Replan will not only "compromise the success of the apportionment count" for Congressional representation, but also "severely compromise the quality of the redistricting data" for state and local representation. Louis Decl. ¶ 43; see Thompson Decl. ¶ 23. In fact, it is undisputed that census data is used to redraw district boundaries for federal, state, and local legislatures, and that drawing districts with unequal population can be unlawful. See, e.g. , Westall Decl. ¶¶ 14–29. An undercount from a truncated self-response period, lower-quality NRFU, and rushed data processing all mean that Plaintiffs’ federal, state, and local political representation will be diminished. See, e.g. , Westall Decl. ¶¶ 27 ("[R]esidents in Council Districts with large concentrations of undercounted residents would be denied equal representation."); Soto Decl. ¶ 11 (same); Ellis Decl. ¶ 12 ("An undercount on the 2020 Census will also put me at serious risk of political underrepresentation in the U.S. Congress, and in the Texas legislature."). 3. The Replan will likely degrade census data that Plaintiffs use to deploy services and allocate capital.
The local government Plaintiffs allege that the Replan will degrade granular census data that they rely on to deploy services and allocate capital. "[B]y virtue of the Constitution and the Census Act, it is, of course, the federal government's job to collect and distribute accurate federal decennial census data." New York v. Trump , 485 F.Supp.3d 422, 457 (S.D.N.Y. 2020) (three-judge court), vacated on ripeness grounds , 592 U.S. ––––, 141 S.Ct. 530, 208 L.Ed.2d 365 (2020) (per curiam); see also Departments of Commerce, Justice, and State, The Judiciary, and Related Agencies Appropriations Act, 1998, § 209, Pub. L. No. 105-119, 111 Stat. 2440, 2481 (1997) ("1998 Appropriations Act") (codified at 13 U.S.C. § 141 note) ("Congress finds that ... it is essential that the decennial enumeration of the population be as accurate as possible, consistent with the Constitution and laws of the United States ....").
The degradation of data is thus an informational injury analogous to those that have supported Article III standing. See New York v. U.S. Dep't of Commerce , 351 F. Supp. 3d 502, 611 (S.D.N.Y. 2019) (finding that "degradation in the quality of census data" supported standing), aff'd in part, rev'd in part and remanded sub nom. Dep't of Commerce v. New York , ––– U.S. ––––, 139 S. Ct. 2551, 204 L.Ed.2d 978 (2019) ; see also, e.g. , Fed. Election Comm'n v. Akins , 524 U.S. 11, 21, 118 S.Ct. 1777, 141 L.Ed.2d 10 (1998) (collecting cases finding that "deprivation of information" supports standing); Robins v. Spokeo, Inc. , 867 F.3d 1108, 1114 (9th Cir. 2017) (finding standing partly because a statute, 15 U.S.C. § 1681e(b), requires "follow[ing] reasonable procedures to assure maximum possible accuracy" of information). For instance, King County, Los Angeles, and Harris County all rely on granular census data:
• King County, Washington uses census data to place public health clinics, plan transportation routes, and mitigate hazards. Dively Decl. ¶ 6.
• The City of Los Angeles uses "reliable, precise, and accurate population count data" to deploy the fire department, schedule trash-pickups, and acquire or improve park properties. Westall Decl. ¶ 32.
• Recently, Harris County has used census data "to estimate the impact of COVID-19 to specific communities at a granular level," which has helped the county tailor "communications in multiple languages with audience and age-specific prevention messaging and share information about availability of testing or vaccine sites." Shah Decl. ¶ 7. Inaccurate or incomplete data would "increase risk of misinterpreting the prevalence of the disease in disproportionately impacted communities." Id.
In sum, the Replan's harm to the accuracy of census data will harm Plaintiffs’ concrete uses of the data.
4. Plaintiffs have diverted and may continue to divert resources to mitigate the undercount that will likely result from the Replan.
Plaintiffs have diverted resources to mitigate the undercounting that will likely result from the Replan. This diversion of resources is a concrete injury-in-fact. See, e.g. , Am. Diabetes Ass'n v. U.S. Dep't of Army , 938 F.3d 1147, 1154 (9th Cir. 2019) (finding injury-in-fact where plaintiffs "had altered their resource allocation" that they would have spent on some other organizational purpose). The City of Salinas, Harris County, Black Alliance for Just Immigration, League of Women Voters, and National Urban League detail many examples of diverted resources:
• The City of Salinas already promoted the October 31 deadline "on social media and in thousands of paper flyers." Gurmilan Decl. ¶¶ 11–12. Thus, "some residents who received the City's messaging will fail to respond before the R[eplan] deadline because the City has limited remaining resources to correct what is now misinformation." Id. ¶ 12. Moreover, the City "advertis[ed] for census enumerator job listings because traditional applicant groups like senior citizens have concerns about the risk of catching COVID-19. With fewer enumerators working, every extra day the City ha[d] to use [ ] existing staff to support the count ...." Id. ¶ 13.
• Harris County "participated in over 150 events," including "food distribution events," during which it "announced the October 31, 2020 deadline for the 2020 Census." Briggs Decl. ¶ 12. Consequently, Harris County "expend[ed] additional resources to clear confusion about the last date for self-response during the Census, to ensure that people who have not responded are counted in time." Id. ¶ 16.
• The Black Alliance for Just Immigration already "publicized the October 31 deadline for self-response during digital events between April and July" and diverted resources to publicize the Replan's September 30 deadline. Gyamfi Decl. ¶¶ 13–14.
• The League of Women Voters "had to spend time and financial resources" developing and distributing public education materials on the Replan timeline. Stewart Decl. ¶ 12.
• The National Urban League has similarly had "to divert resources from other programs and projects" to "alleviate the confusion" about the change in deadlines. Green Decl. ¶ 15.
Indeed, the Census Bureau boasted of how its communications program was "more integrated than ever before" with Plaintiffs such as National Urban League. Fontenot Decl. ¶ 40. Mitigating those counterproductive communications and a likely undercount is only harder in the midst of a pandemic. E.g. , M. Garcia Decl. ¶¶ 14–15; Gurmilan Decl. ¶¶ 11–14; Briggs Decl. ¶¶ 11–12, 15–17. The result that Plaintiffs have diverted and may continue to divert resources from their organizational mission to mitigate the effects of the Replan.
5. Defendants’ supposed >99% completion rate fails to render Plaintiffs’ injuries speculative.
Faced with these detailed allegations of four distinct injuries, Defendants make a new argument in their motion to dismiss: that the >99% completion rate for data collection renders Plaintiffs’ injuries speculative. See Mot. at 15–16. Specifically, Defendants argue that Plaintiffs have inadequately alleged a "differential undercount" because at the end of data collection on October 15, 2020, the Census Bureau had "reached completion rates of greater than 99% in every State and in the District of Columbia." Id. at 15. Defendants’ argument is meritless in five ways.
First, Plaintiffs adequately allege that the asserted completion rates are unreliable. Detailed allegations in the SAC and public reporting cast doubt on the completion rates. Below are just some problems with how the Bureau reached an asserted completion rate of >99%, according to the SAC:
• The Bureau was short over 80,000 enumerators during the data collection period. SAC ¶ 376. Yet the Bureau made up for this shortfall by claiming a 53% increase in enumerator efficiency in the midst of a pandemic. Id. ¶ 378. Driving this "impossible jump in enumerator productivity" was "a massive decrease in in-person visits"; "changed information requirements"; and "enumerators being told to act improperly and cut corners." Id. ¶ 379.
• The decrease in in-person visits led to the "dramatically increased" use of telephone calls, proxies, and administrative records. Id. ¶ 382. Several Census Field Supervisors ("CFSs") reported that households were marked as "completed" without explanation or closed for "political" reasons. Id. ¶¶ 382, 384.
• Many enumerators were instructed to falsify data to inflate the completion rate. For example, a CFS in San Francisco was verbally instructed to take shortcuts that violated the Bureau's manuals. Id. ¶ 405.
As for public reporting, the Associated Press has reported allegations by 12 enumerators that they were "instructed to make up answers about households where they were unable to get information." Mike Schneider, Census Bureau denies fake data allegations by census workers (Nov. 10, 2020), https://apnews.com/article/indiana-only-on-ap-census-2020-a4f3d96a3118f39774fb495b52b4b10a. Text messages from a census supervisor have corroborated similar allegations. See Mike Schneider, Texts: US census manager told counters to use fake answers (Nov. 24, 2020), https://apnews.com/article/us-census-manager-counters-fake-answers-b0ecdec4d6fa4fe3a0989e5702c528ce. Thus, whether the completion rate is reliable warrants discovery and briefing, not dismissal of the SAC.
Second, even if the Bureau truly reached 99% completion in each State, Plaintiffs likely still suffer a differential undercount within states. For instance, Plaintiffs allege that the Bureau's internal documents show "extremely low completion rates in tribal lands." SAC ¶ 368. Those documents also allegedly show that 24 census offices had low completion rates as of September 28, 2020. Id. Nothing so far suggests that the Bureau was able to close the gap in completion rates between September 28 and October 15, 2020.
Third, this Court's Injunction Order is what made possible Defendants’ asserted completion rate. Recall that the Bureau's internal documents insisted that "[i]n order to achieve an acceptable level of accuracy, at least 99% of Housing Units in every state must be resolved." DOC_10275–76. On September 30, 2020—the day on which data collection would have ended but for the Injunction Order—16 States and the District of Columbia were still under 99% completion. See U.S. Census Bureau, 2020 Census Housing Unit Enumeration Progress by State (Oct. 1, 2020 report), https://2020census.gov/content/dam/2020census/news/daily-nrfu-rates/nrfu-rates-report-10-01.pdf. Thus, Defendants would have missed the completion rate requirement despite Associate Director Fontenot's declarations under penalty of perjury that they would not.
Specifically, in his September 4, 2020 declaration, Associate Director Fontenot averred that by September 30, 2020, "the Census Bureau intends to meet a similar level of household responses as in prior censuses, meaning that we will resolve 99% of the cases in each state. " Fontenot Decl. ¶ 65, ECF No. 81-1 (emphasis added). However, on September 17, 2020, Associate Director Fontenot stated to the Census Scientific Advisory Committee ("CSAC") that he "did not know whether Mother Nature would allow us to meet the September 30 date." ECF No. 196-1 ¶ 14. Associate Director Fontenot cited "major West Coast fires and air quality issues that have accelerated since September 11, and the current impacts of Hurricane Sally across the states of Louisiana, Mississippi, Alabama, the Florida panhandle area, parts of Georgia, and South Carolina." Id. ; accord Census Bureau, Census Scientific Advisory Committee Fall Virtual Meeting at 1:36:55–1:37:10 (Sept. 17, 2020), https://www.census.gov/about/cac/sac/meetings/2020-09-meeting.html ("It is not within my vision capabilities to be able to forecast -- That's an unknown at this time ... and I really can't project whether Mother Nature will let us finish.").
Nonetheless, in his September 22, 2020 declaration, Associate Director tried to disavow his inability to forecast whether the Bureau could meet the September 30, 2020 deadline. Associate Director Fontenot averred that, despite Mother Nature, the Bureau still had an "urgent need to revert to our planned completion strategies [under the Replan] to meet the statutory deadline." Id. ¶ 15.
If the Census Bureau had in fact reverted to "planned completion strategies," the Census Bureau would have missed the completion rate requirement detailed above. Even so, Defendants now argue—with little sense of irony—that although Plaintiffs never had standing to begin with, the Injunction Order was so effective that it mooted this case.
Fourth, even if the Bureau collected enough data to hit a high completion rate, the Bureau's severely streamlined data-processing may harm Plaintiffs. Opp'n at 7. As the Bureau's former chief data scientist has averred, "[i]nevitably, streamlining will degrade quality, because the systems and procedures that have been developed and tested over the decade are designed to be both necessary and sufficient." Louis Decl. ¶ 32, ECF No. 36. Current Bureau advisors have echoed this concern too. As the Bureau's Census Scientific Advisory Committee reported on November 12, 2020, "adequate de-duplication procedures for college students, retirees, and others require additional time, not less." Memo. from Allison Plyer, Chair, Census Sci. Advisory Comm. to Steven Dillingham, Dir., U.S. Census Bureau, Recommendations and Comments to the Census Bureau at 3 (Nov. 12, 2020). Yet the current rush to process data will likely undercount (in relative terms) hard-to-count groups that comprise a disproportionate share of Plaintiffs’ members and residents.
Lastly, Defendants fail to adequately address data degradation injury (from lower quality census data that Plaintiffs use for various purposes) and diverted resource injury (from resources spent to mitigate the undercount). Defendants merely reassert that these are speculative self-inflicted harms. Mot. at 17. However, the unrebutted evidence in this case thus far shows that the Replan will "significantly reduce the accuracy" of census data and has caused Plaintiffs to divert resources to mitigate that harm. Injunction Order at 58; see also, e.g. , Louis Decl. ¶ 44 ("truncation of the time for field operations and data curation ... will severely compromise the quality of the census data"); Thompson Decl. ¶ 5 ("grave and material consequences for the 2020 Census").
In sum, the completion of data collection fails to deprive Plaintiffs of standing. Rather, Plaintiffs are still likely to lose federal funds; political representation; accurate data; and diverted resources as a result of the Replan. 6. Plaintiffs’ injuries are fairly traceable to the Replan and redressable by the Court.
The above harms are "concrete, particularized, and actual or imminent." Dep't of Commerce v. New York , 139 S. Ct. at 2565 (quoting Davis v. Fed. Election Comm'n , 554 U.S. 724, 733, 128 S.Ct. 2759, 171 L.Ed.2d 737 (2008) ). They are also "fairly traceable to the defendant's challenged behavior; and likely to be redressed by a favorable ruling.’ " Id. (quoting Davis , 554 U.S. at 733, 128 S.Ct. 2759 ). As the Supreme Court stressed last year, "Article III ‘requires no more than de facto causality.’ " Id. at 2566 (quoting Block v. Meese , 793 F.2d 1303, 1309 (D.C. Cir. 1986) (Scalia, J.)). Defendants’ conduct need not be "the very last step in the chain of causation." Bennett v. Spear , 520 U.S. 154, 169, 117 S.Ct. 1154, 137 L.Ed.2d 281 (1997).
Here, Plaintiffs’ theory of standing rests "on the predictable effect of Government action on the decisions of third parties"—that is, the predictable harms of accelerating census deadlines and curtailing key operations, without warning, after months of publicly operating under a plan tailored to COVID-19. Id. The relief Plaintiffs seek would redress those harms. Specifically, Plaintiffs ask the Court to grant the following relief:
• Declare that Defendants’ promulgation of the Replan, and corresponding revocation of the COVID-19 Plan is unconstitutional under the Enumeration Clause, and unlawful under the Administrative Procedure Act.
• Vacate the Replan, thereby reinstating the COVID-19 Plan.
• Enjoin Defendants from implementing or effectuating the Replan or its constituent parts and enjoin Defendants from unlawfully interfering with the implementation or effectuation of the COVID-19 Plan or its constituent parts.
SAC ¶¶ 486–88. Enjoining Defendants and reinstating the COVID-19 Plan, for instance, would prevent Defendants from trying to "accomplish five months of data processing in ten weeks"—a rush that threatens to render the apportionment numbers "constitutionally and statutorily infirm." SAC ¶ 420. Courts have granted similar relief in other census litigation. See, e.g. , Dep't of Commerce v. U.S. House of Representatives , 525 U.S. at 328–34, 119 S.Ct. 765 (affirming injunction against the planned use of statistical sampling to prevent apportionment harms, among others); New York v. United States Dep't of Commerce , 351 F. Supp. 3d at 675 (issuing injunction to prevent "the loss of political representation and the degradation of information"). Thus, Plaintiffs seek relief that redresses their injuries—injuries which are traceable to the Replan.
In response, Defendants again argue that "the relief Plaintiffs seek—violation of [statutory] deadlines by court-ordered reinstatement of the COVID-19 Plan—is not redress the Court can provide." Mot. at 11 (citing SAC ¶¶ 487–88). Defendants are incorrect in three ways. First, the Ninth Circuit has already held that this Court can afford Plaintiffs’ relief "[i]f the Bureau meets the December 31 [statutory] date by using procedures that violate any accuracy requirement embedded in the Enumeration Clause, or proceeds in an arbitrary and capricious manner." National Urban League II , 977 F.3d at 781. Specifically, "existing data can be reprocessed." Id. This remedy would avoid the problem of "requir[ing] the government to continue to ignore the December 31 timeline for completing the tabulation." Id. at 780 (emphasis in original).
Second, given the case schedule, the Court would not issue a remedy until after the December 31 deadline has passed. See ECF No. 357 at 2 (setting summary judgment hearing on February 18, 2021, and bench trial on March 19, 2021). Thus, no remedy would require Defendants to violate the December 31 deadline. Defendants might have overlooked this point because the Court set the case schedule after Defendants filed their motion to dismiss.
Third, Defendants will miss the December 31 deadline no matter what the Court does. On November 30, 2020, Defendants represented to the Supreme Court that "we are not currently on pace to send the report to the President by the year-end statutory deadline." Oral Arg. Tr. at 6:16–7:3, Trump v. New York , No. 20-366 (U.S. Nov. 30, 2020).
All told, Plaintiffs suffer injuries in fact that are fairly traceable to the Replan and redressable by the relief Plaintiffs seek. Plaintiffs thus still have Article III standing.
C. Plaintiffs’ claims are ripe for adjudication.
After months of litigation—including appeals to the Ninth Circuit and the Supreme Court—Defendants now argue that this case is not ripe for adjudication. Specifically, Defendants invoke constitutional and prudential ripeness. Defendants’ constitutional ripeness argument merely echoes their standing argument and fails on the same grounds detailed above. See Mot. at 18 ("[F]or the same reasons that Plaintiffs have not sufficiently alleged injury-in-fact ... [Plaintiffs’] claims fail to satisfy constitutional ripeness requirements.").
As for prudential ripeness, Defendants make two meritless arguments. First, Defendants argue that "[e]valuating Plaintiffs’ claims of inaccuracy and disproportionate undercount at this juncture is not an inquiry fit for judicial decision." Mot. at 18. Supreme Court precedent forecloses Defendants’ argument. For instance, in last year's litigation about the 2020 Census's citizenship question, the Supreme Court adjudicated Enumeration Clause and APA claims without waiting for undercount-related harms to materialize. See New York , 139 S. Ct. at 2565–66. This Court follows suit for the Enumeration Clause and APA claims here. Similarly, in adjudicating a challenge to the 2000 Census, the Supreme Court held that "the case [was] ripe for review" even though the case had been filed two years before the 2000 Census concluded. Dep't of Commerce v. U.S. House of Representatives , 525 U.S. 316, 328, 119 S.Ct. 765, 142 L.Ed.2d 797 (1999). There, the Supreme Court relied on a statute that creates a cause of action for "any person aggrieved by the use of any statistical method ... in connection with the [ ] census." Id. (quoting 1998 Departments of Commerce, Justice, and State, the Judiciary, and Related Agencies Appropriations Act, § 209, 111 Stat. 2481). Here, Plaintiffs also challenge statistical methods in connection with the census. See, e.g. , SAC ¶¶ 279–88 ("The Replan Did Not Account for Federal Statistical Guidelines").
Second, Defendants argue that "there is no conceivable hardship to Plaintiffs from withholding consideration of their claims." Mot. at 18. Not so. The longer Plaintiffs must wait for potential relief, the more harm they will suffer. Even under the Court's current expedited case schedule, there is no time for delay. The earliest Plaintiffs could receive relief is almost certainly after congressional apportionment. The summary judgment hearing is set for February 18, 2021, and a bench trial is set for March 11, 2021. By either date, presumably Defendants will have determined the results of the 2020 Census.
Moreover, the schedule leaves only three to six weeks for the Court to render judgment before the April 1, 2021 redistricting deadline. Granting judicial relief after redistricting would be more difficult than granting relief before, as this Court has held and as Defendants have represented to the Supreme Court. See, e.g. , ECF No. 372 (discussing "the difficulty of granting relief after the April 1, 2020 deadline for the Secretary's transmittal of redistricting data to the states"). Defendants have specifically represented to the Supreme Court that the 2020 Census must conclude by April 2021 because dozens of states have "contingent redistricting deadlines." Defs. Reply in Support of App. for a Stay at 11, Ross v. National Urban League, ––– U.S. ––––, 141 S.Ct. 18, 208 L.Ed.2d 169 (2020). It is thus illogical for Defendants to argue that this suit should be dismissed entirely as unripe—and refiled by Plaintiffs only months later.
The Supreme Court's recent decision in Trump v. New York is not to the contrary. 592 U.S. ––––, 141 S.Ct. 530, 208 L.Ed.2d 365 (2020) (per curiam); see also ECF No. 394 ("In evaluating Defendants’ motion to dismiss, the Court should consider today's Supreme Court decision in Trump v. New York [ ]."). The New York Court held that the lawsuit against the President's July 21, 2020 Memorandum on Excluding Illegal Aliens from the Apportionment Base Following the 2020 Census is "not suitable for adjudication at this time." id. at 537. The Supreme Court reasoned that in New York , "[u]nlike other pre-apportionment challenges, the Secretary has not altered census operations in a concrete manner that will predictably change the count." Id. at 536. The New York Court also reasoned that "the [ New York ] dispute involves the apportionment process, which remains at a preliminary stage." Id.
The instant case, by contrast, challenges concrete alterations to census operations that will predictably change the count. These concrete alterations include, for example, the Replan's dramatic compression of time for data collection and data processing. Self-response compressed from 33.5 weeks to 29 weeks, with the deadline advancing from October 31 to September 30. Fontenot Decl. ¶ 100. NRFU compressed from 11.5 weeks to 7.5 weeks, with the deadline advancing from October 31 to September 30. Data processing was halved from 26 weeks to 13 weeks, with the deadline advancing from April 30, 2021 to December 31, 2020.
Moreover, as just discussed, the earliest that Plaintiffs could receive relief is February 18, 2021, the date of the summary judgment hearing. At that point, the apportionment process will no longer be "at a preliminary stage." New York , 592 U.S. ––––, 141 S.Ct. at 536. Rather, apportionment will likely be complete. In sum, Plaintiffs’ claims are ripe for adjudication.
D. The Replan still constitutes agency action.
Defendants’ three remaining arguments are against the reviewability of Plaintiffs’ APA claims, not Plaintiffs’ Enumeration Clause claim. All three arguments rehash what this Court and the Ninth Circuit have already rejected. To start, Defendants again argue that the Replan is not reviewable because it is not a discrete "agency action." Mot. At 20. They thus claim that Plaintiffs’ suit is "an improper, programmatic attack on the Bureau's efforts to conduct the 2020 Census." Id. The Court and the Ninth Circuit disagree. See National Urban League II , 977 F.3d at 776 (holding that Defendants "do[ ] not have a strong likelihood" of success on their agency action argument). The Replan is agency action.
"The bite in the phrase ‘final action’ ... is not in the word ‘action,’ which is meant to cover comprehensively every manner in which an agency may exercise its power." Whitman v. Am. Trucking Associations , 531 U.S. 457, 478, 121 S.Ct. 903, 149 L.Ed.2d 1 (2001) (citations omitted). Thus, agency action is broadly defined to include "the whole or part of an agency rule, order, license, sanction, relief, or the equivalent or denial thereof, or failure to act." 5 U.S.C. § 551(13). Each word in that definition has its own expansive definition. A "rule," for example, includes "the whole or a part of an agency statement of general or particular applicability and future effect designed to implement, interpret, or prescribe law or policy or describing the organization, procedure, or practice requirements of an agency." Id. § 551(4).
To be sure, a reviewable agency action must be one that is "circumscribed" and "discrete." Norton v. S. Utah Wilderness All. , 542 U.S. 55, 62–63, 124 S.Ct. 2373, 159 L.Ed.2d 137 (2004). This requirement "precludes [a] broad programmatic attack" on an agency's operations. Id. at 64, 124 S.Ct. 2373. Defendants thus analogize this case to NAACP v. Bureau of the Census , 945 F.3d 183 (4th Cir. 2019). Mot. at 21.
Unlike in their opposition to the preliminary injunction motion, Defendants do not analogize this lawsuit to Lujan v. Nat'l Wildlife Fed'n , 497 U.S. 871, 110 S.Ct. 3177, 111 L.Ed.2d 695 (1990). The Court and the Ninth Circuit explained why National Wildlife was inapposite. See Injunction Order at 31–32 (distinguishing National Wildlife); Nat'l Urban League II , 977 F.3d at 776 (same).
In NAACP , the plaintiffs brought a challenge in 2018 to the census "methods and means," and "design choices." NAACP , 945 F.3d at 186. The NAACP plaintiffs challenged as insufficient the numbers of enumerators, the networks of area census offices, the Bureau's plan to rely on administrative records, and partnership program staffing. Id. at 190. The Fourth Circuit found that " ‘[s]etting aside’ one or more of these ‘choices’ necessarily would impact the efficacy of the others, and inevitably would lead to court involvement in ‘hands-on’ management of the Census Bureau's operations." Id. (citing S. Utah Wilderness All. , 542 U.S. at 66–67, 124 S.Ct. 2373 ). In concluding that there was not final agency action, the Fourth Circuit emphasized that its holding was "based on the broad, sweeping nature of the allegations that the plaintiffs have elected to assert under the APA." Id. at 192.
NAACP is inapposite for two reasons. First, the relief Plaintiffs seek here would not "inevitably [ ] lead to court involvement in ‘hands-on’ management of the Census Bureau[ ]." Id. at 191. Plaintiffs do not ask the Court to manage the Bureau's day-to-day operations or to enforce free-floating standards of "sufficiency." See NAACP , 945 F.3d at 191 (quoting claims of "insufficient network of area census offices," "insufficient partnership program staffing," "insufficient testing of ‘new protocols,’ " and more). Rather, Plaintiffs challenge the Defendants’ failure to consider important aspects of the problem and lack of reasoned explanation for the Bureau's change in position. Reply at 14. See, e.g. , Motor Vehicle Mfrs. Ass'n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co. , 463 U.S. 29, 52, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983) (finding that agency's explanation for rescission was not the product of reasoned decisionmaking); Encino Motorcars, LLC v. Navarro , ––– U.S. ––––, 136 S. Ct. 2117, 2126, 195 L.Ed.2d 382 (2016) (setting aside agency's "change in position" for lacking reasoned explanation).
Second, the Replan is a circumscribed, discrete agency action. Indeed, Defendants treated the Replan accordingly. Defendants named it the "Replan" or "Replanned Operational Schedule." E.g. , DOC_10276 (version of August 3, 2020 slide deck identified as key by the parties); DOC_8929 (July 30, 2020 email from Barbara LoPresti, Chief of the Decennial Information Technology Division, to senior officials discussing "this proposed replan"); DOC_10066 (email thread titled "Replan" with senior officials); DOC_11918 (August 3, 2020 email to the Chief of Staff for the Deputy Secretary of Commerce with subject "Revised Replan Deck").
The Secretary directed the Bureau to develop the Replan. See, e.g. , August 3 Press Release, ECF No. 37-1 ("directed by the Secretary"). In response to the Secretary's direction, the Bureau presented the Replan to the Secretary in a single slide deck. See, e.g. , DOC_10276. The Secretary made a single explicit decision to adopt the Replan. Fontenot Decl. ¶ 85. Census Bureau Director Dillingham announced the Replan in a single press release on August 3, 2020. ECF No. 37-1. Defendants consistently treated the Replan as a circumscribed, discrete agency action. As the Ninth Circuit explained, "[u]nlike in NAACP [ ], Plaintiffs challenge the decisionmaking process that went into the decision in the Replan to greatly accelerate the census process over the COVID-19 Plan, not specific ‘design choices’ within that plan." National Urban League II , 977 F.3d at 776.
Again, as Justice Scalia stated: "[t]he bite in the phrase ‘final action’ ... is not in the word ‘action,’ which is meant to cover comprehensively every manner in which an agency may exercise its power. It is rather in the word ‘final.’ " Whitman , 531 U.S. at 478, 121 S.Ct. 903 (citations omitted). It is to that finality requirement that the Court now turns.
E. The Replan still constitutes final agency action.
Defendants again argue that even if the Replan were agency action, "it is not ‘final’ agency action that is subject to judicial review under § 704." Mot. at 22. "To maintain a cause of action under the APA, a plaintiff must challenge ‘agency action’ that is ‘final.’ " Wild Fish Conservancy v. Jewell , 730 F.3d 791, 800 (9th Cir. 2013) (citing Norton , 542 U.S. at 61–62, 124 S.Ct. 2373 ).
An agency's action is final if two conditions are met. First, the action "must mark the consummation of the agency's decisionmaking process—it must not be of a merely tentative or interlocutory nature." Bennett , 520 U.S. at 177–78, 117 S.Ct. 1154. Second, the action "must be one by which ‘rights or obligations have been determined,’ or from ‘which legal consequences will flow.’ " Id. (quoting Port of Boston Marine Terminal Assn. v. Rederiaktiebolaget Transatlantic , 400 U.S. 62, 71, 91 S.Ct. 203, 27 L.Ed.2d 203 (1970) ). Five years earlier, the Supreme Court found that the same two requirements applied in a census case. Franklin , 505 U.S. at 797, 112 S.Ct. 2767 (the central question "is  whether the agency has completed its decisionmaking process, and  whether the result of that process is one that will directly affect the parties."). Courts should take a " ‘pragmatic’ approach" to finality. U.S. Army Corps of Engineers v. Hawkes Co., Inc. , ––– U.S. ––––, 136 S. Ct. 1807, 1815, 195 L.Ed.2d 77 (2016) (quoting Abbott Laboratories v. Gardner , 387 U.S. 136, 149, 87 S.Ct. 1507, 18 L.Ed.2d 681 (1967) ).
The Replan is final agency action for purposes of APA review because the Replan meets both criteria, each of which the Court addresses in turn. Indeed, the Ninth Circuit has already held that Defendants’ argument to the contrary is likely meritless. See National Urban League II , 977 F.3d at 777 (analyzing finality).
In Hawkes Co. , the Supreme Court expressly reserved whether an agency action that satisfies only the first condition—consummation of the agency's decisionmaking process—can still be final. 136 S. Ct. at 1813 n.2. The Supreme Court did not reach that question in Hawkes Co. because the agency action under review "satisfie[d] both prongs of Bennett ." Id. Similarly, the Replan satisfies both prongs. Thus, the Court need not decide whether the first condition alone would suffice to constitute a "final" agency action.
1. The Census Bureau completed its decisionmaking process: Defendants have adopted and implemented the Replan.
As to the first factor of final agency action, which is "whether the agency has completed its decisionmaking process," Franklin , 505 U.S. at 797, 112 S.Ct. 2767, the Replan marks the consummation of the Bureau's and Department of Commerce's decisionmaking process because the Replan is "not subject to further agency review." Sackett v. EPA. , 566 U.S. 120, 127, 132 S.Ct. 1367, 182 L.Ed.2d 367 (2012) ; see also Hawkes , 136 S. Ct. at 1813–14 (holding that an agency action was final because the determination was "typically not revisited"); Fairbanks North Star Borough v. U.S. Army Corps of Engineers , 543 F.3d 586, 593 (9th Cir. 2008) (holding that an agency's action was final where "[n]o further agency decisionmaking on the issue can be expected"). The Secretary made an explicit decision to adopt the Replan. August 3 Press Release; see Fontenot Decl. ¶ 85. The Bureau has implemented the Replan. No further agency decisionmaking will be conducted on the Replan.
Norton v. Southern Utah Wilderness Alliance , a decision cited by Defendants, is readily distinguishable from the instant case. See Mot. at 20 (citing Norton , 542 U.S. at 61–62, 124 S.Ct. 2373 ). In Norton , the Supreme Court found that the plaintiffs’ challenges to the Bureau of Land Management's land use plans failed. The Norton Court reasoned that the plans were not a "legally binding commitment" that were enforceable under the APA. 542 U.S. at 72, 124 S.Ct. 2373. Specifically, the plaintiffs claimed that BLM "failed to comply with certain provisions in its land use plans," which "describe[ ], for a particular area, allowable uses, goals for future condition of the land, and specific next steps." 542 U.S. at 59, 67, 124 S.Ct. 2373. The Federal Land Policy and Management Act of 1976 "describes land use plans as tools by which ‘present and future use is projected. ’ " Id. at 69, 124 S.Ct. 2373 (emphasis in original) (quoting 43 U.S.C. § 1701(a)(2) ).
Thus, the Norton Court observed that "[t]he implementing regulations make clear that land use plans are a preliminary step in the overall process of managing public lands—designed to guide and control future management actions and the development of subsequent, more detailed and limited scope plans for resources and uses." Id. (emphasis added). As a result, "a land use plan is not ordinarily the medium for affirmative decisions that implement the agency's ‘project[ions].’ " Id. (quoting 43 U.S.C. § 1712(e) ). Similarly, "the regulation defining a land use plan declares that a plan ‘is not a final implementation decision on actions which require further specific plans, process steps, or decisions under specific provisions of law and regulations.’ " Id. at 69–70, 124 S.Ct. 2373. In sum, by contrast to a "final" agency action, the type of land use plan challenged by the Norton plaintiff "is generally a statement of priorities; it guides and constrains actions, but does not (at least in the usual case) prescribe them." Id. at 71, 124 S.Ct. 2373.
Here, the Replan is not a "preliminary step" toward deciding the Census schedule. Nor is the Replan a "statement of priorities" that merely "guides and constrains actions." See id. at 69, 71, 124 S.Ct. 2373. Instead, the Replan constitutes a commitment to terminate the collection of data, analyze that data, and report "[t]he tabulation of total population" to the President by December 31, 2020. 13 U.S.C. § 141(b).
Moreover, termination of data collection is practically irreversible. In his September 5, 2020 declaration, Defendants’ own declarant, Associate Director Fontenot, requested that if the Court enjoins Defendants, the Court do so earlier than later because it is difficult to rehire field staff who have been terminated:
Lack of field staff would be a barrier to reverting to the COVID Schedule were the Court to rule later in September. The Census Bureau begins terminating staff as operations wind down, even prior to closeout. Based on progress to date, as is standard in prior censuses, we have already begun terminating some of our temporary field staff in areas that have completed their work. It is difficult to bring back field staff once we have terminated their employment. Were the Court to enjoin us tomorrow we would be able to keep more staff on board than were the Court to enjoin us on September 29, at which point we will have terminated many more employees.
Fontenot Decl. at ¶ 98. Since that September 5 declaration, data collection has in fact stopped, and there is no indication it will resume before the Secretary reports apportionment figures to the President.
In sum, the Replan provides an accelerated end date for data collection (September 30, 2020, but for the Injunction Order) and truncated data processing. Data collection stopped on October 15, 2020, and accelerated data processing is well underway. The Replan is thus the completion of Defendants’ decisionmaking process on how the 2020 Census will be conducted.
2. The Replan directly affects the parties.
As to the second factor of final agency action, which is whether an agency action "will directly affect the parties," the Replan certainly does affect the parties and will continue to do so. Franklin , 505 U.S. at 797, 112 S.Ct. 2767 ; see also Bennett , 520 U.S. at 177–78, 117 S.Ct. 1154 (holding that, "[a]s a general matter," a final action "must be one by which ‘rights or obligations have been determined,’ or from which ‘legal consequences will flow’ " (citation omitted)). The Court analyzes the Replan's effect on the Plaintiffs and Defendants then distinguishes Defendants’ main case, Franklin v. Massachusetts .
a. The Replan's undercount will directly affect and harm Plaintiffs.
The Replan "will directly affect" Plaintiffs and result in "legal consequences." Franklin , 505 U.S. at 797, 112 S.Ct. 2767 ; Bennett , 520 U.S. at 177–78, 117 S.Ct. 1154. Specifically, the Replan will directly affect Plaintiffs in three ways: (1) by undercounting hard to count populations; (2) barring governmental Plaintiffs’ constituents and organizational Plaintiffs’ members from participating in the 2020 Census after October 15, 2020; and (3) exposing those same people to violations of federal law and fines.
First, the Replan will likely undercount hard to count populations in the decennial census. This undercount necessarily affects the Secretary's "tabulation of total population by States" and the President's apportionment calculations, which "must be based on decennial census data alone." New York , 485 F.Supp.3d at 468-69 (discussing text, legislative history, and the Executive's longstanding understanding of 13 U.SC. § 141(a) and 2 U.S.C. § 2a(a) ). As the Supreme Court recently confirmed, "[t]he President in turn must transmit to Congress a ‘statement showing the whole number of persons in each State, excluding Indians not taxed, as ascertained ’ under the census. " New York , 592 U.S. ––––, 141 S.Ct. at 534 (emphasis added) (quoting 2 U.S.C. § 2a(a) ). In other words, the Replan will likely result in an undercount in both the numbers that the Secretary reports to the States and the numbers that the President—who must draw on "decennial census data"—reports to Congress.
That undercount, as discussed in the Court's standing analysis above, injures Plaintiffs in legally cognizable ways. For instance, an undercount harms the "crucial representational rights that depend on the census," Dep't of Commerce v. New York , 139 S. Ct. at 2569, and deprives local government Plaintiffs of federal funds to which they are entitled, cf. City of Kansas City, Mo. v. U.S. Dep't of Hous. & Urban Dev. , 861 F.2d 739, 745 (D.C. Cir. 1988) (discussing procedural rights arising under Community Development Block Grants, which at least King County and Los Angeles receive). These harms and others will last through 2030, if not later. Congress has determined as much by finding that:
the decennial enumeration of the population is a complex and vast undertaking, and if such enumeration is conducted in a manner that does not comply with the requirements of the Constitution or laws of the United States, it would be impracticable for the States to obtain, and the courts of the United States to provide, meaningful relief after such enumeration has been conducted.
1998 Appropriations Act, § 209(a)(8), 111 Stat. at 2480–81. Thus, because the Replan will likely result in an inaccurate enumeration, the Replan is an action from which legal consequences will flow.
Second, the Replan barred people who seek to participate in the Census—such as governmental Plaintiffs’ constituents and organizational Plaintiffs’ members—from participating after the early end of data collection on October 15, 2020. See Sackett , 566 U.S. at 126, 132 S.Ct. 1367 (holding that an agency action determined rights and obligations of property owners where it "severely limit[ed] [the owners’] ability to obtain a permit ... from [the agency]"); Alaska, Dep't of Environmental Conservation v. EPA , 244 F.3d 748, 750 (9th Cir. 2001) (holding that an agency action determined rights and obligations where its effect was to halt construction at a mine facility). Under the Replan, data collection would have ended on September 30, 2020. The Injunction Order would have extended data collection to October 31, 2020, but for the Supreme Court's stay. Thus, many people were unable to participate despite their potential reliance on the Census Bureau's previous, widely publicized representations that they could participate until October 31, 2020. For example:
• The League of Women Voters has over 65,000 members across 800 state and local affiliates. Stewart Decl. ¶ 4. Thus, "[w]hen the Census Bureau extended the deadline for counting operations to October 31, 2020," the League of Women Voters "published blog posts advertising the new timeline," "shared numerous letters with [ ] state and local affiliates providing information about the new timeline," and "publicized the deadline in letters and [emails]." Id. ¶ 11.
• The City of Los Angeles is home to about 4 million people. M. Garcia Decl. ¶ 7. The City "conducted a public education campaign publicizing the October 31, 2020 date for self-response." Id. ¶ 14. For example, the City announced the date in bus shelter posters and social media toolkits. Id.
• National Urban League has 11,000 volunteers across 90 affiliates in 37 states. Green Decl. ¶ 4. "[W]hen the Census Bureau announced its extension of the timeline for collecting responses to the 2020 Census, the National Urban league informed all members of the 2020 Census Black
Roundtable that the deadline had become October 31, 2020. The members in turn conveyed to their own networks and constituents, causing a cascading effect." Id. ¶ 14.
Third, the Replan exposed the same people—people who believe that October 31, 2020 was the Census deadline—to fines and violations of federal law. By way of background, the Census Act imposes a "clear legal duty to participate in the decennial census." California v. Ross , 362 F. Supp. 3d 727, 739 (N.D. Cal. 2018) (Seeborg, J.) (citing 13 U.S.C. § 221 ). Specifically, 13 U.S.C. § 221(a) provides that any adult who "refuses or willfully neglects ... to answer, to the best of his knowledge, any of the questions on" the census "shall be fined not more than $100." 13 U.S.C. § 221(a). "[E]ach unanswered question" risks an additional fine. Morales v. Daley , 116 F. Supp. 2d at 809 ; accord United States v. Little , 317 F. Supp. 1308, 1309 (D. Del. 1970) ("Presumably there could be a separate violation for each unanswered question."). The 2020 Census form has nine questions for the first person in a household and seven questions for each additional person. See U.S. Census Bureau, 2020 Census Questionnaire (last revised Mar. 7, 2020), https://www.census.gov/programs-surveys/decennial-census/technical-documentation/questionnaires/2020.html. The resulting liability for "refus[ing] or willfully neglect[ing]" to answer an entire Census questionnaire is thus significant. 13 U.S.C. § 221(a).
Because of the excellent publicizing of the COVID-19 Plan, the Replan increased the risk that people will incur that liability. Before the Replan was announced on August 3, 2020, the Bureau and its partners (such as Plaintiff National Urban League) advertised for months that the deadline for census responses was October 31, not September 30 or October 15, 2020. See supra Section III-B-4. As a result, some people may have refused to respond to the questionnaire—or an enumerator's non-response follow-up—on the misunderstanding that they had until October 31, 2020 to comply. This "increase [in] risk of incurring penalties in a future enforcement proceeding" still "constitute[s] ‘legal consequences’ under Bennett ." Ipsen Biopharmaceuticals, Inc. v. Azar , 943 F.3d 953, 957–59 (D.C. Cir. 2019) (emphasis in original) (holding also that "the agency's exercise of prosecutorial discretion" is not enough to render agency action non-final).
b. The Replan directly affects Defendants by binding them for 10 years to a less accurate tabulation of total population.
For Defendants, the Replan gives rise to legal consequences because it effectively binds Defendants—for the next decade—to a less accurate "tabulation of total population by States" under the "decennial census." 13 U.S.C. § 141(b). The Replan does this by committing Defendants to compressing census self-response from 33.5 weeks to 29 weeks; Non-Response Follow Up ("NRFU") from 11.5 weeks to 7.5 weeks; and data processing from 26 weeks to 13 weeks. See, e.g. , Nat. Res. Def. Council v. EPA , 643 F.3d 311, 319–20 (D.C. Cir. 2011) ("[T]he Guidance binds EPA regional directors and thus qualifies as final agency action.").
The result of this significant compression during this unprecedented pandemic will be inaccuracies in the "tabulation of total population." Inaccuracies in the tabulation harm constitutional and statutory interests. See, e.g. , Evans , 536 U.S. at 478, 122 S.Ct. 2191 (finding a "strong constitutional interest in accuracy"); 1998 Appropriations Act, § 209, 111 Stat. at 2481 ("Congress finds that ... it is essential that the decennial enumeration of the population be as accurate as possible ...."). Those constitutional and statutory harms—and Defendants’ choice of speed over accuracy—will endure until 2030.
A less weighty and more easily revocable constraint on the Government was found final in Hawkes Co. , 136 S. Ct. at 1814. There, an internal memorandum of agreement between two federal agencies provided that the Army Corps of Engineers could issue "jurisdictional determinations" ("JDs") that were generally "binding on the Government" for five years. Id. The Supreme Court held that the JDs were final agency action under Bennett v. Spear even though (1) the JDs could be appealed and "revisited," see id. at 1813–14 ; and (2) the JDs’ source of authority, the memorandum of agreement, never went through notice and comment and was represented as non -binding by the United States. See id. at 1817 (opinions of Kennedy, J., concurring; and Ginsburg, J., concurring in part and concurring in the judgment). By contrast, here (1) Defendants did not waver in their commitment to end data collection and are trying to deliver apportionment figures to the President as soon as possible; and (2) there is no doubt that the Replan will bind the United States to this Census and "tabulation of total population" until 2030.
Defendants’ latest argument against final agency action is inapposite. In their motion to dismiss, Defendants argue that the Replan is not final because "the data collection phase of the 2020 Census has already concluded." Mot. at 22. Defendants’ argument is a non-sequitur. The Replan was final before and is final now. A final agency action does not become non-final after it is implemented. To the contrary, the end of data collection shows that the Replan set legal obligations that Defendants then enforced. Regardless, Plaintiffs will suffer many legal consequences that flow not only from data collection, but also data processing. As Census Bureau officials have stated, "[a]ny effort to concatenate or eliminate processing and review steps to reduce the timeframes will significantly reduce the accuracy of the apportionment counts and the redistricting data products." DOC_8337.
Thus, because the Replan determines rights and obligations and gives rise to legal consequences, the Replan constitutes final agency action.
c. Franklin v. Massachusetts shows why the Replan is final agency action.
To argue that the Replan does not constitute final agency action, Defendants rely on the Supreme Court's decision in Franklin v. Massachusetts , 505 U.S. 788, 112 S.Ct. 2767, 120 L.Ed.2d 636 (1992). Mot. at 22. Franklin concerned the Secretary of Commerce's transmission of the census report to the President. Franklin , 505 U.S. at 797–98, 112 S.Ct. 2767. There, the data presented to the President—the allocation of overseas military personnel to states based on their "home of record"—was still subject to correction by the Secretary. Id. In addition, the President could instruct the Secretary to reform the census. Id. at 798, 112 S.Ct. 2767. The Secretary's report to the President thus was a "moving [target]" or a "tentative recommendation," rather than a "final and binding determination." Id. It carried "no direct consequences for the reapportionment." Id. Based on these characteristics, the transmission of the census report was not final agency action. Id. at 798, 112 S.Ct. 2767.
Franklin underscores why the Replan constitutes final agency action. The Replan is neither a "tentative recommendation" nor a decision that will be reviewed by a higher official. Id. Rather, the Secretary directed the Bureau to develop the Replan on July 29, 2020 and approved the Replan on August 3, 2020. Moreover, as a practical matter, no time remains for agency reconsideration. The Replan's field operations ended on October 15, 2020 after the Supreme Court granted Defendants’ application for a stay. See U.S. Census Bureau, Census Bureau Statement on 2020 Census Data Collection Ending (Oct. 13, 2020).
The Replan also has "direct consequences for the reapportionment." Id. The Replan determines when data collection will end—past which people can no longer participate in the census—and solidifies an undercount that will carry through to Congressional reapportionment, federal funding, and more for a decade. By contrast, in Franklin , the data the Secretary reported could have had zero effect. The President could have "reform[ed] the census" and allocated already-counted servicemembers not by "home of record," but by "legal residence," "last duty station," or no "particular State[ ]." Id. at 792, 794, 112 S.Ct. 2767 ; see also U.S. House of Reps. v. U.S. Dep't of Commerce , 11 F. Supp. 2d at 93 (distinguishing Franklin on the same ground).
In any event, "[e]ven in the [ Franklin ] Court's view, the Secretary's report of census information to recipients other than the President would certainly constitute ‘final agency action.’ " Franklin , 505 U.S. at 815 n.14, 112 S.Ct. 2767 (Stevens, J., concurring in part and concurring in the judgment). That is because only the President may order the Secretary "to reform the census, even after the data are submitted to him." Id. at 798, 112 S.Ct. 2767. Data recipients such as the states can do no such thing. Accordingly, the Secretary's reporting of "counts as they are used for intra-state redistricting and for federal fund allocation ... is final agency action for purposes of APA review." City of New York v. U.S. Dep't of Commerce , 822 F. Supp. 906, 918–19 (E.D.N.Y. 1993) (emphasis in original) (challenging guidelines that led Secretary not to adjust undercount), vacated on non-APA grounds , 34 F.3d 1114 (2d Cir. 1994), rev'd sub nom. Wisconsin v. City of New York , 517 U.S. at 12 n.7, 116 S.Ct. 1091 (noting that "[plaintiffs] did not appeal the District Court's treatment of their statutory claims" to the Second Circuit). Plaintiffs here likewise challenge the Replan's undercount as it will be used in intra-state redistricting and federal fund allocation.
Last year's citizenship question cases further underscore why the Replan is final agency action. In those cases, the United States conceded that adding the citizenship question to the census questionnaire constituted final agency action. See New York , 351 F. Supp. 3d at 645 ; Kravitz v. Dep't of Commerce , 336 F. Supp. 3d 545, 566 n.13 (D. Md. 2018). There is no reason that a memorandum announcing the addition of a question would mark the agency "complet[ing] its decisionmaking process" and "directly affect[ing] the parties," Franklin , 505 U.S. at 797, 112 S.Ct. 2767, but the Replan would not. In both cases, the Secretary directed the development of and adopted the Replan; the Bureau viewed the Secretary's decision as binding; and the decision directly affects the parties. In sum, the Replan is final agency action.
F. The Replan is not committed to agency discretion by law.
Defendants’ last argument on reviewability is that the administration of the census—including the Replan—is "committed to agency discretion by law." 5 U.S.C. § 701(a)(2). The Court disagrees.
The APA creates a "strong presumption favoring judicial review of administrative action." Mach Mining, LLC v. EEOC , 575 U.S. 480, 489, 135 S.Ct. 1645, 191 L.Ed.2d 607 (2015). One exception includes those actions that are "committed to agency discretion by law." 5 U.S.C. § 701(a)(2). However, courts have read this exception quite narrowly. This exception encompasses situations where Congress explicitly precludes review, or " ‘those rare circumstances where the relevant statute is drawn so that a court would have no meaningful standard against which to judge the agency's exercise of discretion.’ " Weyerhaeuser , 139 S. Ct. at 370 (quoting Lincoln v. Vigil , 508 U.S. 182, 191, 113 S.Ct. 2024, 124 L.Ed.2d 101 (1993) ). This latter exception has generally been limited to "certain categories of administrative decisions that courts traditionally have regarded as committed to agency discretion ... such as a decision not to institute enforcement proceedings ... or a decision by an intelligence agency to terminate an employee in the interest of national security." Dep't of Commerce , 139 S. Ct. at 2568 (citations and quotation marks omitted) (citing Heckler v. Chaney , 470 U.S. 821, 831–32, 105 S.Ct. 1649, 84 L.Ed.2d 714 (1985), and Webster v. Doe , 486 U.S. 592, 600–01, 108 S.Ct. 2047, 100 L.Ed.2d 632 (1988) ).
Department of Commerce v. New York controls. There, the Supreme Court concluded that "[t]he taking of the census is not one of those areas traditionally committed to agency discretion." 139 S. Ct. at 2568. Collecting case law, the Supreme Court noted that "courts have entertained both constitutional and statutory challenges to census-related decisionmaking." Id. (citing, e.g. , Carey , 637 F.2d at 839, in which the Second Circuit concluded that the Bureau's decision not to use "Were You Counted" forms or to compare census records with records of Medicaid-eligible people "was not one of those ‘rare instances’ where agency action was committed to agency discretion"); see also City of Los Angeles v. U.S. Dep't of Commerce , 307 F.3d 859, 869 n.6 (9th Cir. 2002) (rejecting argument that the Bureau's decision not to adopt statistically adjusted population data was committed to agency discretion by law). The Supreme Court explained that there were meaningful standards against which to judge the taking of the census, including the Census Act, which requires that the agency "conduct a census that is accurate and that fairly accounts for the crucial representational rights that depend on the census and the apportionment." Id. at 2568–69 (quoting Franklin , 505 U.S. at 819–20, 112 S.Ct. 2767 (Stevens, J., concurring in part and concurring in judgment)).
Here, Plaintiffs challenge the Replan—a set of deadlines for "the taking of the census." Id. at 2568. Plaintiffs’ claims, like those in Department of Commerce v. New York , arise under the Enumeration Clause and the APA. Here too, the Census Act provides a meaningful standard against which to judge Defendants’ action. The Replan's change in deadlines affects the accuracy of the enumeration, as did the decision to omit certain records in Carey or reinstate the citizenship question in New York . Accordingly, the Replan is not committed to agency discretion.
In sum, Defendants’ motion to dismiss is meritless.
G. Staying this case is unwarranted.
The Court now turns to Defendants’ motion to stay this case. In determining whether to grant a stay, the Court weighs three factors: (1) the possible damage to the non-moving party, (2) "the hardship or inequity which a party may suffer in being required to go forward," and (3) "the orderly course of justice." CMAX , 300 F.2d at 268 ; see also Lockyer , 398 F.3d at 1110 (applying Landis factors). None of the factors supports staying this case.
First, Plaintiffs would likely suffer irreparable harm from a stay. As detailed in the standing analysis above, see supra Section III-B, Plaintiffs rely on accurate census data for federal funds, political representation, and the allocation of limited resources for a decade. Thus, the longer Plaintiffs must wait for potential relief, the more harm they will suffer. Under the Court's current case schedule, there is no time for delay. The earliest Plaintiffs could receive relief is likely after congressional apportionment because the summary judgment hearing is on February 18, 2021, and the bench trial begins on March 19, 2021. Both dates are well after January 11, 2021—the deadline for the President's transmittal of congressional apportionment figures to Congress. See 2 U.S.C. § 2a(a) (setting deadline).
Second, Defendants will not suffer "hardship or inequity" without a stay. CMAX , 300 F.2d at 268. For one, Defendants’ stay motion rests on an outdated premise. Defendants’ stay motion assumes that the Court set a schedule of "expedited litigation with the goal of having the Court enter judgment before the end of December." ECF No. 355 at 8. That expedited schedule, Defendants argue, "would be extremely disruptive." Id. Yet after Defendants filed their stay motion, the Court rejected Plaintiffs’ proposed schedule. ECF No. 357. The Court will not enter judgment until February 18, 2021 at the earliest. Thus, Defendants’ claim of "extreme[ ] disrupti[on]" fails to account for the additional months that Defendants now have to manage this lawsuit.
For another, Defendants fail to show hardship or inequity under the current case schedule. "[B]eing required to defend a suit, without more, does not constitute a ‘clear case of hardship or inequity’ within the meaning of Landis ." Lockyer , 398 F.3d at 1112. Defendants’ claimed hardships are simply the burden of "being required to defend a suit." Id. Specifically, Defendants assert that the lawsuit will "burden[ ] employees at a time when their work is most crucial." ECF No. 355 at 8. Lockyer holds that this burden fails to support a stay.
Moreover, the Court has imposed significant limits on discovery that reduce any hardship on Defendants. For instance, although the Federal Rules of Civil Procedure (the "Federal Rules") permit 10 depositions per side, the Court cut that number nearly in half, permitting only 6 depositions. See Fed. R. Civ. P. 30(a)(1) ; ECF Nos. 357 (case management order), 372 (Order to Compel). Similarly, the Federal Rules permit 25 interrogatories per side, but the Court cut that number by 60%, permitting only 10 interrogatories. See Fed. R. Civ. P. 33(a)(1) ; ECF No. 357. The Federal Rules also permit an unlimited number of Requests for Admission ("RFAs") and Requests for Production ("RFPs"), but the Court permitted only 25 RFAs and 25 RFPs per side in the instant case. See Fed. R. Civ. P. 34, 36 ; ECF No. 357. Thus, Defendants fail to make a "clear case of hardship or inequity" that would result from litigating this case. Lockyer , 398 F.3d at 1112.
Lastly, the orderly course of justice militates against a stay. The orderly course of justice is "measured in terms of the simplifying or complicating of issues, proof, and questions of law which could be expected to result from a stay." CMAX , 300 F.2d at 268. Moreover, the Ninth Circuit has "repeatedly admonished district courts not to delay trial preparation to await an interim ruling on a preliminary injunction." California v. Azar , 911 F.3d 558, 583–84 (9th Cir. 2018). Here, Defendants ask this Court to do what the Ninth Circuit has admonished against. Specifically, Defendants ask the Court to wait "until Defendants’ appeal of the preliminary injunction is resolved and the Bureau completes the 2020 Census." ECF No. 355 at 5. The grounds for this request are stale. They are the same reviewability and ripeness grounds that this Court and the Ninth Circuit have rejected. See Section III-A through F, supra. Thus, the orderly course of justice fails to support Defendants’ stay motion. In sum, none of the stay factors support staying this case.
For the foregoing reasons, the Court DENIES Defendants’ motion to dismiss the Second Amended Complaint and Defendants’ motion for stay of proceedings. Because fact discovery closes on January 7, 2021, Defendants shall file an answer to the Second Amended Complaint by December 30, 2020.