Timothy H. EDGAR; Richard H. Immerman; Melvin A. Goodman; Anuradha Bhagwati; Mark Fallon, Plaintiffs - Appellants, v. Avril D. HAINES, in her official capacity as Director of National Intelligence; David Cohen, in his official capacity as Director of the Central Intelligence Agency; Lloyd J. Austin, III, in his official capacity as Secretary of Defense; Paul M. Nakasone, in his official capacity as Director of the National Security Agency, Defendants - Appellees. Professor Jack Goldsmith; Professor Oona A. Hathaway; The Reporters Committee for Freedom of the Press, Amici Supporting Appellant.
Five former employees of our Nation's security agencies who, during their employment, had clearances for access to classified and sensitive information, commenced this action against the Central Intelligence Agency (CIA), the Department of Defense (DoD), the National Security Agency (NSA), and the Office of the Director of National Intelligence (ODNI), facially challenging the agencies’ requirements that current and former employees give the agencies prepublication review of certain materials that they intend to publish. These prepublication review requirements allow the agencies to redact information that is classified or otherwise sensitive to the national security. The employees alleged in their complaint that this prepublication review — which is implemented through “regimes” of policies, regulations, and individual employee agreements — violates their free speech rights guaranteed by the First Amendment and their rights under the Due Process Clause of the Fifth Amendment. Specifically, they alleged that the agencies’ regimes “fail to provide former government employees with fair notice of what they must submit,” “invest executive officers with sweeping discretion to suppress speech[,] and fail to include procedural safeguards designed to avoid the dangers of a censorship system.”
The district court, in a thorough and well-reasoned opinion, granted the defendant agencies’ motion to dismiss, holding that their prepublication review regimes were “reasonable” measures to protect sensitive information and thereby did not violate the plaintiffs’ First Amendment rights. The court held further that the regimes were not unduly vague under the Fifth Amendment because they adequately informed authors of the types of materials they must submit and established for agency reviewers the kinds of information that can be redacted.
We agree with the district court and affirm.
Information related to national security has, since World War I, been graded according to sensitivity under a classification system. See Dep't of Navy v. Egan, 484 U.S. 518, 527, 108 S.Ct. 818, 98 L.Ed.2d 918 (1988); see also Daniel Patrick Moynihan et al., Report of the Commission on Protecting and Reducing Government Secrecy, S. Doc. No. 105-2, app. A (“Secrecy: A Brief Account of the American Experience”) (1997). And security agencies have, over the years, adopted policies and regulations to protect classified information from public disclosure. They have also required various employees to sign agreements, as a condition of employment or as a condition for receiving access to classified information, requiring the employees to follow the agencies’ policies and regulations. Currently, information that is subject to classification includes “military plans, weapons systems, or operations”; “foreign government information”; “intelligence activities”; “foreign activities of the United States”; and “vulnerabilities or capabilities of ․ infrastructures ․ relating to the national security”; as well as a few other categories of a similarly sensitive nature. Exec. Order No. 13,526, Classified National Security Information, 75 Fed. Reg. 707, 709 (Dec. 29, 2009).
Under current classifications, information that, if disclosed, “reasonably could be expected to cause damage to the national security” is classified as “Confidential”; information the disclosure of which “reasonably could be expected to cause serious damage to the national security” is classified as “Secret”; and information that, if disclosed, “reasonably could be expected to cause exceptionally grave damage to the national security” is classified as “Top Secret.” Exec. Order No. 13,526, 75 Fed. Reg. at 707–08 (emphasis added). In addition, when information “concern[s] or [is] derived from intelligence sources, methods[,] or analytical processes” that require protection “within formal access control systems,” it may be further designated as “Sensitive Compartmented Information,” or “SCI.” Intelligence Community Directive 703, Protection of Classified National Intelligence, Including Sensitive Compartmented Information § 2 (June 21, 2013).
Disclosing information involving national security can be detrimental to the vital national interest, and courts have recognized that the government has “a compelling interest in protecting ․ the secrecy of [such] important” information. Snepp v. United States, 444 U.S. 507, 509 n.3, 100 S.Ct. 763, 62 L.Ed.2d 704 (1980) (per curiam). As a consequence, agencies involved in intelligence and national security currently have in place, through policies and regulations, a range of practices and procedures designed to protect against the inappropriate disclosure of information related to national security. One such practice and procedure is “prepublication review,” which requires current and former employees to submit materials intended for publication to their agencies to enable the agencies to redact, in advance of publication, classified or otherwise sensitive information. This prepublication review process — which is the subject of the plaintiffs’ challenge here — relies on the agency's judgment about what is sensitive and detrimental to the national security and therefore must be redacted, rather than on the employee's independent judgment. This is because the agency has a “broader understanding of what may expose classified information and confidential sources.” Id. at 512, 100 S.Ct. 763.
Under the prepublication review process adopted by each of the defendant agencies, current and former employees are required to submit to their agencies a broad scope of materials that relate to their employment and experience with the agency and that they intend to publish. The agency reviews the materials for classified and sensitive information and, to protect against disclosure of that information, directs that it be redacted, thereby ensuring that the information will not be inadvertently disclosed by the author. The details of the process for each defendant agency are as follows.
The CIA: CIA Agency Regulation 13-10, Agency Prepublication Review of Certain Material Prepared for Public Dissemination (June 25, 2011), provides that employees, former employees, “and others who are obligated by CIA secrecy agreement” must “submit for prepublication review” “any written, oral, electronic, or other presentation intended for publication or public dissemination, whether personal or official, that mentions CIA or intelligence data or activities on any subject about which the author has had access to classified information in the course of his employment or other contact with the” CIA. Id. § 2(b)(1), (e)(1). The CIA reviews proposed publications “solely to determine whether [they] contain[ ] any classified information.” Id. § 2(f)(2). And “[a]s a general rule, the [CIA] will complete prepublication review ․ within 30 days of receipt of the material.” Id. § 2(d)(4). The regulation explains, however, that while “short, time-sensitive submissions ․ will be handled as expeditiously as practicable,” “[l]engthy or complex submissions may require a longer period of time for review.” Id. Authors dissatisfied with the initial reviewer's decisions can appeal within the CIA. Id. § 2(h)(1). Consistent with this policy, CIA employees must also sign an agreement as a condition of employment, agreeing “to submit for review by the [CIA] any writing or other preparation in any form, including a work of fiction, which contains any mention of intelligence data or activities, or contains any other information or material that might be based on” classified information or information the author knows is “in the process of a classification determination.” The agreement explains that prepublication review is meant to give the CIA “an opportunity to determine whether the information or material ․ contains any” classified information the employee received in the course of employment, which the employee, by signing the agreement, has “agreed not to disclose.” The term of the agreement is indefinite.
The DoD: Current, former, and retired DoD employees, contractors, and military service members who have had access to DoD information and facilities must submit for prepublication review “[a]ny official DoD information intended for public release that pertains to military matters, national security issues, or subjects of significant concern to the DoD.” DoD Instruction 5230.09, Clearance of DoD Information for Public Release § 1.2(b) (Jan. 25, 2019); Frequently Asked Questions for Security and Policy Reviews of Articles, Manuscripts, Books, and Other Media Prior to Public Release, DoD (Mar. 2012), https://perma.cc/5AH3-S3RV. “Official DoD information” is defined as “information that is in the custody and control of the DoD, relates to information in the custody and control of the DoD, or was acquired by DoD personnel as part of their official duties or because of their official status within DoD.” DoD Instruction 5230.09, glossary § G.2. And prepublication review is defined as “[t]he process by which information ․ is examined ․ for compliance with established national and DoD policies and to determine whether it contains any classified, export-controlled[,] or other protected information.” Id. DoD policy explains that “[t]he public release of official DoD information is limited only as necessary to safeguard information requiring protection in the interest of national security or other legitimate government interest.” Id. § 1.2(d). For former employees, prepublication review is meant “to ensure that information” they “intend to release to the public does not compromise national security as required by their nondisclosure agreements.” Id. § 1.2(g). DoD regulations also provide that “security review protects classified information, controlled unclassified information, or unclassified information that may individually or in aggregate lead to the compromise of classified information or disclosure of operation security.” DoD Instruction 5230.29, Security and Policy Review of DoD Information for Public Release, enclosure 3 § 1 (Apr. 14, 2017). The DoD advises authors to submit papers and articles “at least 10 working days” before the anticipated publication date and manuscripts and books “at least 30 working days” in advance. Id. enclosure 3 § 3(a)(2), (4). Dissatisfied authors are authorized to appeal within the DoD. Id. enclosure 3 § 4(b).
The NSA: Current and former NSA employees acting in a private capacity may publish materials using information that is “unclassified and approved for public release,” but they must submit proposed materials for prepublication review where “compliance with” that requirement “is in doubt.” NSA/CSS Policy 1-30, Review of NSA/CSS Information Intended for Public Release, §§ 2, 6(b), 10(a) (May 12, 2017) (cleaned up); see also id. § 30 (defining prepublication review as “[t]he overall process to determine that information proposed for public release contains no protected information”); 50 U.S.C. § 3605(a) (providing, subject to certain exceptions, that no law “shall be construed to require the disclosure of the organization or any function of the National Security Agency, or any information with respect to the activities thereof, or of the names, titles, salaries, or number of the persons employed by such agency”). The NSA sets for itself a 25-day goal for reviewing a proposed publication. NSA/CSS Policy 1-30 § 6(b)(7). Dissatisfied authors are authorized to appeal within the NSA. Id. § 7.
The ODNI: ODNI regulations require current and former ODNI employees to submit any “publication that discusses the ODNI, the [Intelligence Community], or national security” to the ODNI for prepublication review. ODNI Instruction 80.04, Rev. 2, ODNI Pre-Publication Review of Information to be Publicly Released §§ 4, 6 (Aug. 9, 2016). “The goal of prepublication review is to prevent the unauthorized disclosure of information, and to ensure the ODNI's mission and the foreign relations or security of the U.S. are not adversely affected by publication.” Id. § 3. The ODNI thus reviews submitted materials “to safeguard sensitive intelligence information and prevent its unauthorized publication.” Id. § 6; see 50 U.S.C. § 3024(i)(1) (“The Director of National Intelligence shall protect intelligence sources and methods from unauthorized disclosure”). The ODNI's policy is to “complete a review of non-official publication requests no later than 30 calendar days from the receipt of the request, as priorities and resources allow.” ODNI Instruction 80.04 § 6(C)(2)(b). Dissatisfied authors are authorized to appeal within the ODNI. Id. § 6(E). Consistent with this policy, ODNI employees also sign an ODNI-specific nondisclosure agreement as a prerequisite for accessing classified information that is materially identical to the CIA's secrecy agreement.
All four agencies also authorize referrals of proposed publications to other agencies that have equities at stake in a proposed disclosure.
In addition to these agency-specific policies, the plaintiffs’ complaint describes various nondisclosure agreements that employees are required to sign as a condition of accessing classified or sensitive information. Thus, when an employee signs Standard Form 312, entitled “Classified Information Nondisclosure Agreement,” the employee agrees to “never divulge classified information to anyone” unless the employee has “officially verified that the recipient has been properly authorized ․ to receive it” or has “been given prior written notice of authorization ․ that such disclosure is permitted.” The employee also agrees “to comply with laws and regulations that prohibit the unauthorized disclosure of classified information.” And when an employee signs Standard Form 4414, entitled “Sensitive Compartmented Information Nondisclosure Agreement,” which applies to employees who need access to SCI, the employee agrees similarly to “never divulge” SCI “to anyone who is not authorized to receive it without prior written authorization.” The employee also agrees to “submit for security review,” by the agency that granted the employee SCI access, “any writing or other preparation in any form, including a work of fiction, that contains or purports to contain any SCI or description of activities that produce or relate to SCI or that [the employee] ha[s] reason to believe are derived from SCI.” Both of these nondisclosure forms impose obligations that apply during employment “and at all times thereafter.” Other general or agency-specific agreements referred to in the complaint, such as Form 313 and DD Form 1847-1, contain similar provisions.
In common, all four defendant agencies require — whether by policy, regulation, agreement, or a combination of them — that all current and former employees submit to the agency materials that they intend to publish to give the agency the opportunity to require redaction of classified or sensitive information. This prepublication review process may be analogized to a funnel. At the top end, a broad scope of materials intended for publication is called for and entered into the review process — materials that might contain classified or sensitive information. And at the bottom end, only a narrow scope of materials is selected for redaction — materials that actually contain classified or sensitive information.
The plaintiffs are five former employees of three of the four defendant agencies. Because they alleged that the prepublication review process at these agencies is facially unconstitutional, their personal experiences with the publication of agency-related materials in the past — which are detailed at some length in the complaint — are mostly relevant only to determine the plaintiffs’ standing and the ripeness of their action (which the agencies challenge in this case).
Plaintiff Timothy Edgar was an ODNI employee from 2006 to 2013 and held a Top Secret/SCI clearance. In October 2016, Edgar submitted a manuscript for his book Beyond Snowden: Privacy, Mass Surveillance, and the Struggle to Reform the NSA to the ODNI for review. The ODNI referred the manuscript to both the CIA and the NSA for additional review, and review was completed in January 2017. Edgar alleged that some of the required redactions “related to events that had taken place, or issues that had arisen, after [he] had left government” and that others “related to facts that were widely discussed and acknowledged though perhaps not officially confirmed.” He did not, however, challenge the mandated redactions because he did not want to delay publication of the book and because he wanted to maintain “a good relationship with reviewers at the ODNI.” Edgar alleged that he plans to continue writing in this field and “anticipates submitting at least some” publications for review, but he also alleged that the review requirement “has dissuaded him from writing some pieces that he would have otherwise written[ ] and has caused him to write others differently than he would otherwise have written them.”
Plaintiff Richard Immerman was an ODNI employee from 2007 to 2009 and held a Top Secret/SCI clearance. In January 2013, he submitted a manuscript for the book The Hidden Hand: A Brief History of the CIA to the ODNI's prepublication review office. The ODNI referred the manuscript to the CIA, and review was completed in July 2013. Immerman alleged that some of the proposed redactions “related to information that had been published previously by government agencies”; that other redactions related to public information; and that several others “related to events that had taken place, or issues that had arisen, after [he] had left government.” Immerman appealed those redactions within the ODNI, and the ODNI “informed him that he could publish a significant portion of the” redacted text, and the CIA agreed that “some of the [proposed] redactions were unnecessary.” Immerman thereafter published his book, which included “roughly eighty percent of the material that the agencies had originally redacted.” Immerman alleged that he plans to submit more articles and books in this field and that he “would publish more” if it were not for the “burdens and uncertainties associated with prepublication review.”
Plaintiff Melvin Goodman was a CIA employee from 1966 to 1990 and held a Top Secret/SCI clearance. Upon joining the CIA, he signed the standard secrecy agreements. Since leaving the CIA, he “has published nine books and has submitted each manuscript to the CIA for prepublication review.” While the review process “typically took less than two months,” “the CIA took eleven months to review a manuscript of his latest book, Whistleblower at the CIA.” Goodman “believes that all of the” CIA's “changes ․ were intended to spare the agency embarrassment, not to protect classified information.” Moreover, Goodman alleged that some of the redactions concerned “widely reported aspects of U.S. government policy.” As Goodman also alleged, he “intends to submit” for review “those portions of any future manuscripts that deal with intelligence matters,” but he worries that the CIA “will demand that he redact material unwarrantedly ․ and that the delay associated with prepublication review will jeopardize his book contracts and render his publications less relevant to quickly evolving public debates.”
Plaintiff Anuradha Bhagwati is a former Marine Corps officer who was cleared to receive Secret information. She recently published Unbecoming: A Memoir of Disobedience, “a memoir that centers on her confrontation of misogyny, racism, and sexual violence in the military, as well as her advocacy on related issues after leaving the Marines.” Bhagwati, however, did not submit that book for prepublication review and “has no plans to submit any future work to prepublication review.” But she alleged that she remains concerned that the DoD might sanction her for failing to submit her work for review.
Plaintiff Mark Fallon is a former employee of the DoD and other agencies who held Top Secret and Top Secret/SCI clearances. In January 2017, Fallon submitted a manuscript of his book Unjustifiable Means to the DoD's prepublication review office, and review was completed in August 2017. Fallon alleged that the proposed redactions were “intended to protect the CIA from embarrassment” and that “[s]ome of them related to material that had been published in unclassified congressional reports.” Fallon also alleged that “he is unsure whether he is willing to embark on writing on another book” and “has declined offers to author op-eds and write articles on topics of public concern” because of “potential delays and unjustified objections by the agency.” He has, however, recently “submitted numerous shorter works” and a book chapter for review.
While the plaintiffs have alleged their personal circumstances, they do not challenge the application of prepublication review to any specific work. Rather, their complaint alleged that facially the prepublication review “regime” of each agency is “a far-reaching system of prior restraints that suppresses a broad swath of constitutionally protected speech, including core political speech, by former government employees.” After describing the regimes in some detail, their complaint concluded:
Defendants’ prepublication review regimes violate the First Amendment because they invest executive officers with sweeping discretion to suppress speech and fail to include procedural safeguards designed to avoid the dangers of a censorship system.
Defendants’ prepublication review regimes are void for vagueness under the First and Fifth Amendments because they fail to provide former government employees with fair notice of what they must submit for prepublication review and of what they can and cannot publish, and because they invite arbitrary and discriminatory enforcement.
For relief, the plaintiffs sought a declaratory judgment that the defendants’ “prepublication review regimes violate the First and Fifth Amendments to the Constitution”; an injunction prohibiting the defendants “from continuing to enforce [their] prepublication review regimes against Plaintiffs, or any other person”; and costs and attorneys fees.
The defendant agencies filed a motion to dismiss the complaint under Federal Rules of Civil Procedure 12(b)(1) and 12(b)(6), contending (1) that the plaintiffs lacked standing as required by Article III of the Constitution; (2) that the plaintiffs’ claims were unripe; and (3) that, in any event, the plaintiffs failed to state a claim under either the First or Fifth Amendment.
The district court rejected the agencies’ arguments for dismissal based on a lack of standing or ripeness. The court held that the plaintiffs had standing because they plausibly alleged that the defendant agencies’ prepublication review regimes had “a chilling effect on protected speech.” Edgar v. Coats, 454 F. Supp. 3d 502, 523, 525–27 (D. Md. 2020). And it ruled that the plaintiffs’ claims were ripe because they were challenging policies to which they “are currently subject ․ that they reasonably allege require them to self-censor.” Id. at 530. But the court granted the agencies’ motion to dismiss on the merits, concluding that the plaintiffs had failed to state a plausible claim. The court explained that prepublication review regimes are not classic prior restraints and are instead consistent with the First Amendment so long as they are “reasonable.” Id. at 530–32 (quoting Snepp, 444 U.S. at 509 n.3, 100 S.Ct. 763). It found that “Plaintiffs have failed to demonstrate that the regimes do not meet” that “low threshold.” Id. at 537. The court also rejected the plaintiffs’ vagueness claim, noting that the plaintiffs’ primary issue with the regimes’ submission requirements “is their breadth rather than any difficulties Plaintiffs have in understanding what they require.” Id. at 539. The court then parsed the agencies’ separate prepublication review regimes and concluded that they “appear to set out reasonable limitations and guidance” for reviewers. Id. at 541.
From the district court's order of dismissal dated April 16, 2020, the plaintiffs filed this appeal.
We address first our jurisdiction, which the defendant agencies have challenged in arguing that the plaintiffs lack Article III standing to bring their action and that the issues are not ripe for adjudication. The district court rejected both arguments, and for substantially the same reasons given by the district court, we affirm its rulings on these issues.
The defendant agencies contend first that the district court erred in finding standing. On that issue, the court concluded that the plaintiffs “plausibly alleged that features of the [prepublication review] regimes result in a chilling effect on the exercise of First Amendment rights” and therefore “have made a sufficient showing of an injury in fact to proceed.” Edgar, 454 F. Supp. 3d at 527. The defendants argue, however, that the “Plaintiffs fail[ed] to show that the challenged features of defendants’ policies would cause any objectively reasonable chill,” as necessary to establish the injury-in-fact element for establishing Article III standing.
Article III's standing requirement centers “on whether the party invoking jurisdiction had the requisite stake in the outcome when the suit was filed.” Davis v. FEC, 554 U.S. 724, 734, 128 S.Ct. 2759, 171 L.Ed.2d 737 (2008). At this stage, a party has such a stake when it is able to plausibly allege “(1) an injury in fact, (2) a sufficient causal connection between the injury and the conduct complained of, and (3) a likelihood that the injury will be redressed by a favorable decision.” Susan B. Anthony List v. Driehaus, 573 U.S. 149, 157–58, 134 S.Ct. 2334, 189 L.Ed.2d 246 (2014) (cleaned up). These requirements are, however, “somewhat relaxed in First Amendment cases,” given that even the risk of punishment could “chill[ ]” speech. Cooksey v. Futrell, 721 F.3d 226, 235 (4th Cir. 2013); see also Sec'y of State of Md. v. Joseph H. Munson Co., 467 U.S. 947, 956, 104 S.Ct. 2839, 81 L.Ed.2d 786 (1984). Thus, “[i]n First Amendment cases, the injury-in-fact element is commonly satisfied by a sufficient showing of ‘self-censorship, which occurs when a claimant is chilled from exercising his right to free expression.’ ” Cooksey, 721 F.3d at 235 (cleaned up) (quoting Benham v. City of Charlotte, 635 F.3d 129, 135 (4th Cir. 2011)). But this chilling effect “must be objectively reasonable.” Benham, 635 F.3d at 135 (cleaned up). In short, while plaintiffs need not show that the government action led them to stop speaking “altogether,” they must show that the action would be “likely to deter a person of ordinary firmness from the exercise of First Amendment rights.” Id. (cleaned up).
Here, the plaintiffs asserted that the vagueness and breadth of the defendants’ prepublication review regimes required them “to submit far more than [they] should be required to submit”; allowed agency officials to “redact material unwarrantedly”; and caused them to write some pieces “differently than [they] would have otherwise written them.” The plaintiffs further alleged that these infirmities, together with the delays created by the defendants’ prepublication review regimes, have “dissuaded [them] from writing some pieces” they “would have otherwise written,” and have made it more difficult to engage in “quickly evolving public debates.”
These are, we conclude, adequate allegations of an “objectively reasonable” chill sufficient to show that the defendants’ prepublication review regimes are “likely to deter a person of ordinary firmness from the exercise of First Amendment rights.” Benham, 635 F.3d at 135 (cleaned up). Importantly, some plaintiffs alleged that they have decided not to write about certain topics because of the prepublication review policies. Such self-censorship is enough “for an injury-in-fact to lie.” Cooksey, 721 F.3d at 236.
The plaintiffs’ allegations also satisfy the causation and redressability elements of the standing inquiry. See Susan B. Anthony List, 573 U.S. at 158, 134 S.Ct. 2334. The chilling of the plaintiffs’ speech was plainly alleged to have been caused by the particular prepublication review regimes at issue here. As the plaintiffs alleged, they would publish more but for those regimes. See Cooksey, 721 F.3d at 238 (“[C]ausation is satisfied where a causal connection between the injury and the conduct complained of ․ is fairly traceable, and not the result of the independent action of some third party not before the court” (cleaned up)). And there is more than “a non-speculative likelihood that th[is] injury would be redressed by a favorable judicial decision.” Id. (cleaned up). A favorable decision on the plaintiffs’ behalf would deem the defendants’ regimes unconstitutional and enjoin the defendants from enforcing them.
Accordingly, we reject the defendant agencies’ argument that the plaintiffs lack Article III standing to challenge the prepublication review regimes.
On ripeness, the defendant agencies argue that the plaintiffs’ claims are “paradigmatically unripe” because they arise “in the absence of a concrete factual dispute.” According to the defendants, courts require a specific application of prepublication review to determine “whether plaintiffs’ treatment has been unfair.” The defendants also contend that requiring the plaintiffs “to litigate their claims in the context of a concrete dispute” would not cause them any material hardship; as they argue, the plaintiffs “who are dissatisfied with the review decisions can challenge them in court.”
“Like standing, the ripeness doctrine originates in the ‘case or controversy’ constraint of Article III.” South Carolina v. United States, 912 F.3d 720, 730 (4th Cir. 2019) (cleaned up). “The question of whether a claim is ripe turns on the ‘fitness of the issues for judicial decision’ and the ‘hardship to the parties of withholding court consideration.’ ” Id. (ultimately quoting Abbott Labs. v. Gardner, 387 U.S. 136, 149, 87 S.Ct. 1507, 18 L.Ed.2d 681 (1967)). Thus, while standing considers who may sue, ripeness considers when they may sue. There is, however, “obvious overlap between the doctrines.” Id. (cleaned up). And “[m]uch like standing, ripeness requirements are also relaxed in First Amendment cases.” Cooksey, 721 F.3d at 240.
The plaintiffs have challenged practices and procedures to which they are currently subject and which, they plausibly alleged, require them to self-censor. These are legal issues for which no “further factual development” is necessary. Va. Soc'y for Hum. Life, Inc. v. FEC, 263 F.3d 379, 390 (4th Cir. 2001), overruled on other grounds by The Real Truth About Abortion, Inc. v. FEC, 681 F.3d 544 (4th Cir. 2012)). And deciding them does not require us to interpret the agencies’ policies and regulations in the “abstract”; we instead are called to decide what conduct the plaintiffs “can engage in without threat of penalty.” Id. Therefore, their claims are fit for judicial review. Moreover, the plaintiffs “will face a significant impediment if we delay consideration of the regulation's constitutionality.” Id. As the plaintiffs allege, they are currently curbing their speech in light of the defendants’ prepublication review regimes. See Cooksey, 721 F.3d at 240 (“First Amendment rights are particularly apt to be found ripe for immediate protection, because of the fear of irretrievable loss” (cleaned up)). Thus, the plaintiffs have adequately demonstrated that refusing to reach their claims would cause them material hardship.
For these reasons, we agree with the district court and conclude that the plaintiffs’ claims are ripe for adjudication.
On the merits, the plaintiffs contend first that the defendant agencies’ prepublication review regimes — consisting of, as they characterize them, a “confusing tangle of contracts, regulations, and policies” — violate their First Amendment rights because the regimes “invest executive officers with sweeping discretion to suppress speech and fail to include procedural safeguards designed to avoid the dangers of a censorship system.” More particularly, they argue that the regimes have overly broad and confusing submission requirements; include “confusing, subjective, and overbroad” review standards that “do not meaningfully limit [officials’] censorship authority”; and lack “any definite deadlines for decisions.”
Addressing first the employment agreements, the complaint alleged that as part of the regimes imposing prepublication review, the defendant agencies require employees to sign one or more forms of nondisclosure agreements “as a prerequisite to accessing classified information.” The complaint describes numerous standard forms, including Form 312, Form 313, Form 4414, a “standard CIA secrecy agreement,” and DD Form 1847-1, all allegedly containing employee promises not to disclose classified or sensitive information without prior authorization. The agreements make clear that this is a continuing obligation, applicable even after the employee leaves the agency. Moreover, some of the agreements, particularly Form 4414, describe the process of submitting intended writings for prepublication review.
No plaintiff has alleged that he or she was coerced into signing any agreement or was under any duress in doing so. Indeed, no plaintiff even contends that the agreements were, as contracts, invalid. They challenged only the agreements’ contribution to the implementation of “prepublication review,” which they contend violates their First Amendment rights as an unlawful prior restraint.
The Supreme Court, however, has already said that such agreements are “not unenforceable as [ ] prior restraint[s].” Snepp, 444 U.S. at 509 n.3, 100 S.Ct. 763. Indeed, the Court has blessed a similar agreement as a “reasonable means for protecting” the government's “compelling interest in protecting both the secrecy of information important to our national security and the appearance of confidentiality so essential to the effective operation of our foreign intelligence service.” Id. And we have held that in signing such nondisclosure agreements, the employee “effectively relinquishe[s] his First Amendment rights” to the sensitive information those agreements protect. Alfred A. Knopf, Inc. v. Colby, 509 F.2d 1362, 1370 (4th Cir. 1975); see also Wilson v. CIA, 586 F.3d 171, 183 (2d Cir. 2009) (“[O]nce a government employee signs an agreement not to disclose information properly classified pursuant to executive order, that employee simply has no first amendment right to publish such information” (cleaned up)); Stillman v. CIA, 319 F.3d 546, 548 (D.C. Cir. 2003) (same).
Accordingly, by voluntarily signing these agreements, the plaintiffs knowingly waived their First Amendment rights to challenge the requirement that they submit materials for prepublication review and the stated conditions for prepublication review. For the most part, that could end the matter. Yet, because the plaintiffs challenge the clarity of the stated conditions and their interpretive scope, as well as the manner in which the defendant agencies have implemented prepublication review, such as its timeliness, we turn to address the challenges that they make.
In challenging prepublication review, the plaintiffs identify four specific aspects that they claim render the defendants’ entire regimes unconstitutional under the First Amendment. First, they contend that the scope of matters subject to prepublication review is too broad, “sweep[ing] in virtually everything that former intelligence agency employees might write about the government.” Second, they contend that the scope of persons subject to the submission requirements is too expansive, applying to “all former employees — not just those who had access to SCI.” Third, they contend that the review standards are “confusing, subjective, and overbroad,” allowing the defendants “to censor information ․ whether or not it was obtained by the author in the course of employment; ․ whether or not its disclosure would actually cause harm; ․ whether or not it is already in the public domain; and ․ whether or not the public interest in its disclosure outweighs the government's interest in secrecy.” And fourth, they contend that the prepublication review process lacks firm or binding deadlines, allowing for inappropriate delays.
At the outset, we reiterate that the plaintiffs are mounting a facial challenge, meaning that their claim is that the policies and regulations are unconstitutional not as applied to their own conduct, but rather, on their face, as they apply to the population generally. United States v. Miselis, 972 F.3d 518, 530 (4th Cir. 2020). Such facial challenges “are disfavored” because they “run contrary to the fundamental principle of judicial restraint that courts should neither anticipate a question of constitutional law in advance of the necessity of deciding it nor formulate a rule of constitutional law broader than is required by the precise facts to which it is to be applied.” Wash. State Grange v. Wash. State Republican Party, 552 U.S. 442, 450, 128 S.Ct. 1184, 170 L.Ed.2d 151 (2008) (cleaned up). Accordingly, facial challenges typically require “a showing that no set of circumstances exists under which the [law] would be valid, i.e., that the law is unconstitutional in all of its applications, or that the statute lacks any plainly legitimate sweep.” Miselis, 972 F.3d at 530 (cleaned up). But given the “fear of chilling protected expression,” id., a facial challenge to a law on the ground that it is overbroad under the First Amendment can be successful “if a substantial number of its applications are unconstitutional, judged in relation to the statue's plainly legitimate sweep,” United States v. Stevens, 559 U.S. 460, 473, 130 S.Ct. 1577, 176 L.Ed.2d 435 (2010) (emphasis added) (cleaned up).
The relevant constitutional standard that we must apply in addressing this facial challenge derives from Snepp. That case concerned the remedy available to the CIA when a former agent, who agreed to prepublication review upon joining the CIA, nonetheless published a book about certain CIA activities without submitting it for prepublication review. Snepp, 444 U.S. at 507–08, 100 S.Ct. 763. The Court held that the agent's profits from the book should be subject to a constructive trust in favor of the CIA. Id. at 509–10, 100 S.Ct. 763. And, as critical here, in conducting its analysis, the Court rejected the agent's argument that the agreement was an unconstitutional prior restraint. It explained that the government can “impos[e] reasonable restrictions on employee activities that in other contexts might be protected by the First Amendment.” Id. at 509 n.3, 100 S.Ct. 763. And the nondisclosure agreement that included prepublication review was, the Court held, a “reasonable means for protecting” the government's “compelling interest in protecting both the secrecy of information important to our national security and the appearance of confidentiality so essential to the effective operation of our foreign intelligence service.” Id.
Snepp’s analysis amounted, at its core, to an application of a reasonableness test that balances “the interests of the [employee], as a citizen, in commenting upon matters of public concern” with “the interest of the [government], as an employer, in promoting the efficiency of the public services it performs through its employees.” Pickering v. Bd. of Educ., 391 U.S. 563, 568, 88 S.Ct. 1731, 20 L.Ed.2d 811 (1968); see also Weaver v. U.S. Info. Agency, 87 F.3d 1429, 1439 (D.C. Cir. 1996) (noting that the Snepp Court “essentially applied Pickering”). And when this reasonableness test is applied to a regulation that operates as a prior restraint on employee speech, the government must show “that the interests of both potential audiences and a vast group of present and future employees in a broad range of present and future expression are outweighed by that expression's ‘necessary impact on the actual operation’ of the Government.” United States v. Nat'l Treasury Employees Union, 513 U.S. 454, 468, 115 S.Ct. 1003, 130 L.Ed.2d 964 (1995) (quoting Pickering, 391 U.S. at 571, 88 S.Ct. 1731).
Because Snepp determined that the government has a “compelling interest” in the secrecy of information important to national security, the question in this case reduces to whether the defendant agencies’ prepublication review regimes are a reasonable and effective means of serving that interest.
First, with respect to the plaintiffs’ argument that the scope of materials subject to prepublication review is overly broad and therefore not reasonable in serving the government's interest, it is true that the defendants’ submission standards do cover a broad range of materials. But this is necessary to serve the government's compelling interest because the aim of prepublication review is, as the parties agree, to prevent the inadvertent disclosure of sensitive information. Thus, the scope of materials subject to review must include materials that might contain, reveal, or confirm classified or sensitive information. And that is what the defendants’ submissions standards do.
The CIA requests all material that “mentions CIA or intelligence data or activities on any subject about which the author has access to classified information.” CIA AR 13-10 § 2(e)(1). The DoD, all material containing “official DoD information ․ that pertains to military matters, national security issues, or subjects of significant concern to the DoD.” DoD Instruction 5230.09 § 1.2(b). The NSA, any material that may not adhere to the NSA's requirement that employees not publish classified information or information not approved for public release. NSA/CSS Policy 1-30 §§ 2, 6(b), 10(a). And the ODNI requires that employees submit any “publication that discusses the ODNI, the [Intelligence Community], or national security.” ODNI Instruction 80.04 § 6.
Distilled to their essence, these submission standards are designed to reach materials that reasonably could reveal classified information or information sensitive to the national security and thus are reasonably tied to the goal of avoiding the inadvertent disclosure of such information. And importantly, the scope of materials subject to review is not the same as the scope of materials that may not be published. The scope of materials for review simply identifies materials that are subject to the process. We conclude that these submission requirements are not overly broad.
Second, with respect to the plaintiffs’ contention that the scope of persons covered by the submission is overly broad, we reject the argument for similar reasons. The requirement — that all current and former employees who have had access to certain types of information are covered by the policy — is reasonably tied, indeed necessary, to the government's interest. This is just another way of ensuring that certain types of information are not inadvertently disclosed. For instance, a low-level employee in a security agency who has received no clearances yet becomes aware of information that, if published, could lead to the disclosure of classified information presents the same interests justifying prepublication review as an employee with proper clearance. Because the scope of persons subject to review is cabined by the definition of the materials subject to review, it is therefore not unreasonable.
Third, the plaintiffs’ argument that the standards for redaction are overly vague and broad is belied by the text of the policies and regulations, all of which are geared to the redaction of classified information, information that is otherwise restricted or could lead to the disclosure of classified information, or information that the agencies are under a statutory requirement to protect. See CIA AR 13-10 § 2(f)(2) (“classified information”); DoD Instruction 5230.09 § 1.2(g) (“information” that “compromise[s] national security” in violation of the employees’ “nondisclosure agreements”); id. glossary § G.2 (“classified, export-controlled or other protected information”); DoD Instruction 5230.29, enclosure 3 § 1 (“classified information, controlled unclassified information, or unclassified information that may individually or in aggregate lead to the compromise of classified information or disclosure of operation security”); NSA/CSS Policy 1-30 §§ 2, 6(b), 10(a) (classified information or information not approved for public release); ODNI Instruction 80.04 § 6 (“sensitive intelligence information”).
While the plaintiffs claim that the scope of redaction authority includes information that was not obtained by the author in the course of his or her employment or information that is already in the public domain, those circumstances do not render unreasonable the criteria focused on classified or otherwise sensitive material. The plaintiffs all enjoyed positions of trust in the government, involving national security, and were granted access to classified or otherwise sensitive information while so employed. By virtue of those positions, the public is likely to view such officials as speaking with authority — indeed it is often because of that authority that former officials engage in public discussions about governmental affairs at all. But, as we have explained, “[i]t is one thing for a reporter or author to speculate or guess that a thing may be so or even, quoting undisclosed sources, to say that it is so; it is quite another thing for one in a position to know of it officially to say that it is so.” Alfred A. Knopf, Inc., 509 F.2d at 1370. That is because an official's repetition of information that is already in the public domain but not yet unclassified, or his speaking on information that is classified but post-dates his time in the respective agency, “lend[s] credence” to that information and could, in the eyes of the public, confirm the existence of such classified information. Id. Such confirmation, of course, can be as good as official disclosure to those who are paying attention.
Fourth and finally, we conclude that the plaintiffs’ argument that the defendants’ policies and regulations fail to establish firm or binding deadlines for the review — thereby unreasonably chilling speech — lacks merit in the circumstances presented. We recognize that a drawn-out process “might delay constitutionally protected speech to a time when its only relevance was to historians.” Weaver, 87 F.3d at 1441 (citing FW/PBS, Inc. v. City of Dallas, 493 U.S. 215, 228, 110 S.Ct. 596, 107 L.Ed.2d 603 (1990)). But considering the policies and regulations facially, as the plaintiffs request, the regimes here fix target timelines for review. Moreover, the plaintiffs’ allegations do not, on the whole, indicate that the agencies failed to abide by these timelines. Instead, the plaintiffs pointed to a few specific book-length manuscripts that the defendants allegedly failed to review in a timely manner. But even if the time periods for those reviews were inappropriately long — something we do not reach — those few allegations do not suffice to find the policies and regulations unconstitutional across the board. See Stevens, 559 U.S. at 472–73, 130 S.Ct. 1577.
At bottom, we conclude that the defendant agencies’ prepublication review regimes are a reasonable means of serving the government's compelling interest in keeping classified or otherwise sensitive information secret, and therefore they do not violate the plaintiffs’ First Amendment speech rights.
The plaintiffs also contend that the defendant agencies’ prepublication review regimes are unconstitutionally vague under the Due Process Clause, as well as the First Amendment, because, as they argue, the regimes “fail to give former employees fair notice of what they must submit for review” and “fail to provide explicit standards for reviewers, thus inviting arbitrary and discriminatory enforcement.”
A fundamental component of due process is that “laws which regulate persons or entities must give fair notice of conduct that is forbidden or required.” FCC v. Fox Television Stations, Inc., 567 U.S. 239, 253, 132 S.Ct. 2307, 183 L.Ed.2d 234 (2012); see also Manning v. Caldwell, 930 F.3d 264, 272 (4th Cir. 2019) (en banc). And a regulation that “fails to provide a person of ordinary intelligence fair notice of what is prohibited, or is so standardless that it authorizes or encourages seriously discriminatory enforcement,” is impermissibly vague and must therefore be invalidated. Fox Television, 567 U.S. at 253, 132 S.Ct. 2307 (cleaned up); see also Manning, 930 F.3d at 272. “These twin concerns of inadequate notice and arbitrary or discriminatory enforcement are especially pronounced” when a regulation implicates speech “because ambiguity inevitably leads citizens to steer far wider of the unlawful zone than if the boundaries were clearly marked, thereby chilling protected speech.” Miselis, 972 F.3d at 544 (cleaned up); see also In re Murphy-Brown, LLC, 907 F.3d 788, 800 (4th Cir. 2018). That said, however, “perfect clarity and precise guidance have never been required even of regulations that restrict expressive activity.” Miselis, 972 F.3d at 544 (quoting United States v. Williams, 553 U.S. 285, 304, 128 S.Ct. 1830, 170 L.Ed.2d 650 (2008)).
The plaintiffs argue first that the defendant agencies’ prepublication review regimes do not give them adequate notice of what must be submitted for review, advancing essentially the same reasons that they advanced for contending that the regimes violate the First Amendment. But in doing so, they focus more particularly on the use of “terms such as ‘relates to,’ ‘pertains to,’ ‘subjects of significant concern,’ and ‘might be based upon,’ ” which they argue are “ambiguous terms” that “force former employees to guess at whether they must submit their speech for review.”
This argument, however, misses the forest for the trees. To be sure, terms such as “pertains to” and “might be based upon” do result in broad submission standards, the exact contours of which could be hazy in the abstract. Indeed, there even may be “close cases” at the extreme edges, but close cases do not make a regulation vague. Williams, 553 U.S. at 306, 128 S.Ct. 1830; see also Bruce & Tanya & Assocs., Inc. v. Bd. of Supervisors, No. 19-1151, ––– Fed.Appx. ––––, ––––, 2021 WL 1854750, at *6 (4th Cir. May 10, 2021) (“[D]ue process demands a measure of clarity, not exactitude”). But crucially, these abstract terms are all anchored to discrete and identifiable categories of information, thereby narrowing the scope of submission in such a way that employees of ordinary intelligence would know what needs to be submitted. See Williams, 553 U.S. at 304, 306, 128 S.Ct. 1830. To take just one example, the DoD requires the submission of materials containing “official DoD information ․ that pertains to military matters, national security issues, or subjects of significant concern to the DoD.” DoD Instruction 5230.09 § 1.2(b) (emphasis added). Given that the goal of prepublication review is to prevent the accidental disclosure of information sensitive to the national security, requiring former employees of national-security and intelligence agencies to submit materials that, for instance, “pertain to” the national security is a sufficiently “sensible basis for distinguishing what” must be submitted for review and what can be published immediately. Minn. Voters Alliance v. Mansky, ––– U.S. ––––, 138 S. Ct. 1876, 1888, 201 L.Ed.2d 201 (2018).
The plaintiffs also argue that the defendants’ “censorship standards” for deciding what to redact “fail to provide ‘explicit standards for those who apply them,’ inviting ‘arbitrary and discriminatory enforcement.’ ” (Quoting Grayned v. City of Rockford, 408 U.S. 104, 108, 92 S.Ct. 2294, 33 L.Ed.2d 222 (1972)). But this argument is largely a repackaging of the plaintiffs’ First Amendment argument, and we reject it for the same reasons, i.e., because all the defendants’ redaction standards are guided by whether material discloses classified information or otherwise sensitive information. Most of the categories of restricted information are binary: Either information is classified or it is not; either it is “controlled” or it is not; and it has either been “approved for public release” or it has not. And the few standards that are not binary provide “meaningful guidance” to reviewers. Manning, 930 F.3d at 275. For example, the DoD can restrict the publication of “unclassified information” only if it “may ․ lead to the compromise of classified information or disclosure of operation security.” DoD Instruction 5230.29, enclosure 3 § 1. In short, the defendants’ review standards “adequately define the range of” information that cannot be published by authors and accordingly provide sufficient guidance to reviewers to prevent arbitrary censorship. Miselis, 972 F.3d at 545.
At bottom, we hold that the defendants’ prepublication review regimes adequately define for authors the types of materials that they must submit for review and adequately establish for reviewers the types of information that cannot be published. Accordingly, they are not unconstitutionally vague.
The national security agencies’ policies and regulations that the plaintiffs challenge here are all directed at ensuring the Nation's security and maintaining security-related secrets, which go to the core of the agencies’ mission. And the plaintiffs’ employment contributing to fulfilling that mission was especially important national service. For this, the plaintiffs can be proud, and the public is grateful.
But the plaintiffs’ special employment carried with it a serious responsibility not to impair the agencies’ work, which could be compromised irreversibly by the inadvertent disclosure of national secrets. While it is understandable that the plaintiffs, as former employees, now wish to share their experiences or, yet more, to comment on public policy as informed by those experiences, doing so in light of their exposure to numerous state secrets is fraught with danger to the national security. And it goes without saying that national security is one of the federal government's overarching responsibilities — one necessary to the protection of the liberties guaranteed by the Constitution — and therefore must be given a high priority. It is thus a compelling interest.
In this case, we conclude that in balancing the effective protection of national security secrets with the speech interests of former employees and the public, we must, as necessary to serve the national interest, require some give in the plaintiffs’ speech interests. And indeed, in the employment agreements that the plaintiffs signed, they freely gave their assent to this.
Taking the defendant agencies’ policies and regulations facially and as a whole, we therefore conclude that the prepublication review regimes established by them do not violate the plaintiffs’ rights under the First and Fifth Amendments. The judgment of the district court is accordingly
NIEMEYER, Circuit Judge:
Affirmed by published opinion. Judge Niemeyer wrote the opinion, in which Judge Keenan and Judge Traxler joined.