Grand Jury Subpoena v.

2017 | Cited 0 times | Ninth Circuit | November 8, 2017



IN RE GRAND JURY SUBPOENA, No. No. 17-16221 16-03-217, D.C. No. 2:17-mc-00036- UNITED STATES OF AMERICA, DJH Appellee,


GLASSDOOR, INC., Movant-Appellant.

Appeal from the United States District Court for the District of Arizona Diane J. Humetewa, District Judge, Presiding

Argued and Submitted October 16, 2017 San Francisco, California

Filed November 8, 2017

Before: Richard C. Tallman and Consuelo M. Callahan, Circuit Judges, and David A. Ezra, * District Judge.

Opinion by Judge Tallman

* The Honorable David A. Ezra, United States District Judge for the District of Hawaii, sitting by designation.



Subpoena / First Amendment Rights

The panel affirmed the district court’s denial of Glassdoor, Inc.’s motion to quash a grand jury subpoena duces tecum that would require Glassdoor to disclose the identifying information of eight users who posted anonymous reviews about another company on its Internet website; and sustained the contempt order entered to enforce it.

Glassdoor argued that complying with the subpoena would violate its users’ First Amendment rights to associational privacy and anonymous speech.

The panel held that to determine whether the subpoena violated the First Amendment, the proper test on the record of this case was the good-faith test the Supreme Court established in Branzburg v. Hayes, 408 U.S. 665 (1972). The panel rejected Glassdoor’s contention that the district court should have applied the compelling-interest test laid out in Bursey v. United States, 466 F.2d 1059 (9th Cir. 1972).

The panel held that because Glassdoor had neither alleged nor established bad faith on the part of the government in its investigation, under Branzburg, enforcement of the subpoena duces tecum to identify potential witnesses in aid of its inquiries did not violate the First Amendment rights of Glassdoor’s uses. The panel

** This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader.


further held that Glassdoor had not shown that any other evidence was necessary to rule on its objection.


Eric D. Miller (argued), Nicola C. Menaldo, and Todd M. Hinnen, Perkins Coie LLP, Seattle, Washington, for Movant-Appellant.

Andrew C. Stone (argued) and Gary M. Restaino, Assistant United States Attorneys; Krissa M. Lanham, Deputy Appellate Chief; Elizabeth A. Strange, Acting United States Attorney; United States Attorney’s Office, Phoenix, Arizona; for Appellee.

Kurt Wimmer and Jadzia Butler, Covington & Burling LLP, Washington, D.C.; Sophia Cope, Electronic Frontier Foundation, San Francisco, California; for Amici Curiae Center for Democracy & Technology, Committee for Justice, Electronic Frontier Foundation, Media Alliance, and Public Participation Project.


TALLMAN, Circuit Judge:

Glassdoor, Inc. appeals the denial of its motion to quash a grand jury subpoena duces tecum that would require Glassdoor to disclose the identifying information of eight users who posted anonymous reviews about another company on its Internet website, Glassdoor argues that complying with the subpoena would violate its users’ First Amendment rights to associational privacy and


anonymous speech. It contends that the district court should have applied the compelling-interest test we laid out in Bursey v. United States, 466 F.2d 1059 (9th Cir. 1972), to determine whether the subpoena violates the First Amendment. The government argues that the good-faith test the Supreme Court established in Branzburg v. Hayes, 408 U.S. 665 (1972), controls.

We agree that on the record before us, Branzburg, which was decided the day before we issued Bursey, supplies the proper test. Because there is no evidence that the grand jury’s investigation of fraud, waste, and abuse by a third party in performing a government contract is being conducted in bad faith, we affirm the denial of the motion to quash, and we sustain the contempt order entered to enforce it.


Glassdoor, Inc. operates, a website where employers promote their companies to potential employees, and employees post reviews of what it’s like to work at their companies. In these reviews, employees rate their employers in a variety of categories and describe workplace environments, salaries, and interviewing practices.

The reviews on are anonymous. But to post reviews, users must first provide Glassdoor with their e-mail addresses, though the addresses do not appear on the site. Before Glassdoor accepts a posting, the contributor is warned his or her information may be disclosed when required by law, either through a subpoena or court order. Glassdoor’s Terms of Use incorporate a Privacy Policy that assures users the company generally “do[es] not disclose . . . individual account or usage data to third parties.” But the


Privacy Policy expressly warns users that Glassdoor “will disclose data if we believe in good faith that such disclosure is necessary . . . to comply with relevant laws or to respond to subpoenas or warrants or legal process served on us.” Finally, Glassdoor’s Terms of Use also inform users that Glassdoor reserves the right “to take appropriate action to protect the anonymity of [its] users against the enforcement of subpoenas or other information requests.” The company is attempting to do that here in the face of an ongoing federal criminal investigation.

An Arizona federal grand jury is investigating a government contractor that administers two Department of Veterans Affairs (VA) healthcare programs. The grand jury is examining whether the subject 1 of its inquiries has committed wire fraud and misused government funds in violation of 18 U.S.C. § 1343 and 18 U.S.C. § 641, respectively.

As of March 2017, current and former employees of the subject company had posted 125 reviews on

1 We use the term “subject” to mean the object of a grand jury’s inquiries prior to the point in an investigation where a subject becomes a “target,” against whom the grand jury is contemplating returning an indictment based on evidence obtained during its investigation. A “subject” is “a person whose conduct is within the scope of the grand jury’s investigation,” while a “target” is “a person as to whom the prosecutor or the grand jury has substantial evidence linking him or her to the commission of a crime and who, in the judgment of the prosecutor, is a putative defendant.” U.S. Attorneys’ Manual, § 9–11.151 (2009). The investigation here is at an early stage and may well result in the grand jury’s decision not to accuse the subject of any criminal acts. We do not identify the subject of these inquiries to protect—to the extent we can while adjudicating this subpoena challenge—the secrecy embodied in Federal Rule of Criminal Procedure 6(e), designed to protect innocent parties who are ultimately never indicted for any crime.


Many of the reviews criticize the subject’s management and business practices. For example, one anonymous employee wrote that it “[m]anipulate[s] the system to make money unethically off of veterans/VA.” Another asserted that “[t]here’s a real disconnect between how this program runs and how the VA thinks the program runs.”

On March 6, 2017, the government served Glassdoor with a subpoena that ordered it to provide the grand jury with “Company Reviews” and associated “reviewer information” for every review of the subject on The requested “reviewer information” included “internet protocol addresses and logs associated with all reviews including date and time of post, username, email address, resume, billing information such as first name, last name, credit card information, billing address, payment history, and any additional contact information available.” The government attached eight “exemplar reviews,” all of which were critical of the subject.

Glassdoor notified the government that it believed “the scope of the request raise[d] issues associated with the First Amendment.” The government agreed to limit its request to the reviewer information associated with just the eight exemplar reviews, and it told Glassdoor the information would enable it “to contact those reviewers as third party witnesses to certain business practices relevant to [its] investigation.” Glassdoor maintained its objection to the subpoena and filed a motion to quash.

The district court denied Glassdoor’s motion. It held that Bursey’s compelling-interest test was inapplicable because the facts of Bursey were distinguishable. Applying Branzburg, it held that Glassdoor had not shown the grand jury investigation was being conducted in bad faith, and it


ordered Glassdoor to respond to the subpoena on pain of contempt.

Glassdoor chose to bring a recalcitrant witness appeal rather than comply with the subpoena. 28 U.S.C. § 1826. The parties stipulated to a judgment of civil contempt and sanctions of $5,000 per day until Glassdoor fully complies by producing the requested information. The district court entered an order in accordance with the stipulation of contumacious conduct and stayed enforcement of the monetary sanctions pending resolution of this appeal. We have jurisdiction under 28 U.S.C. § 1291.


We review the district court’s denial of a motion to quash a grand jury subpoena, as well as the district court’s imposition of contempt sanctions, for abuse of discretion. See In re Grand Jury Subpoena (Mark Torf/Torf Envtl. Mgmt.), 357 F.3d 900 , 906 (9th Cir. 2004) (denial of motion to quash); In re Grand Jury Proceedings, 33 F.3d 1060 , 1061 (9th Cir. 1994) (per curiam) (contempt).

“[M]ixed questions of law and fact contained within the analysis of a civil contempt proceeding” are reviewed de novo. In re M.H., 648 F.3d 1067 , 1070–71 (9th Cir. 2011). The district court’s “underlying factual findings are reviewed for clear error.” Mathews v. Chevron Corp., 362 F.3d 1172 , 1180 (9th Cir. 2004).


Glassdoor argues that the grand jury subpoena violates its users’ First Amendment rights in two ways: it infringes


on their right to associational privacy and their right to anonymous speech. 2

Implicit in the First Amendment is a “right to associate for the purpose of engaging in those activities protected by the First Amendment.” Roberts v. United States Jaycees, 468 U.S. 609 , 618 (1984). Because there is a “vital relationship between freedom to associate and privacy in one’s associations,” in some circumstances, forcing organizations to disclose their members’ identities can infringe on their associational rights. NAACP v. Alabama, 357 U.S. 449 , 462, 466 (1958).

Glassdoor claims that its users constitute an expressive association because they “associate online with other users and fellow employees to discuss the conditions at their jobs.” Therefore, requiring Glassdoor to identify eight of its users violates those users’ right to associational privacy.

Glassdoor’s associational claim is tenuous. The Supreme Court’s expressive-association jurisprudence contemplates groups of people who have associated to advance shared views or “join in a common endeavor,” Jaycees, 468 U.S. at 618, not people who happen to use a common platform to anonymously express their individual views. See, e.g., Boy Scouts of Am. v. Dale, 530 U.S. 640 ,

2 The government does not contest that Glassdoor has third-party standing to assert the rights of its users, but we must satisfy ourselves that Glassdoor has standing. See Bd. of Nat. Res. of State of Wash. v. Brown, 992 F.2d 937 , 945 (9th Cir. 1993). We are satisfied here because Glassdoor has established an injury in fact of its own, it has a sufficiently close relationship to its users, and its users would face “genuine obstacles” to the assertion of their own putative right to anonymity. See McCollum v. Cal. Dep’t. of Corr. & Rehab., 647 F.3d 870 , 879 (9th Cir. 2011).


649–50 (2000) (holding that the Boy Scouts is an expressive association because it is an organization with the “mission” of “instill[ing] values in young people”); Jaycees, 468 U.S. at 622 (“According protection to collective effort on behalf of shared goals is especially important in preserving political and cultural diversity . . . .”) (emphasis added); NAACP, 357 U.S. at 460 (“Effective advocacy of both public and private points of view . . . is undeniably enhanced by group association.”).

Given the nature of, it is difficult to see its users as an expressive association like the Jaycees, the Boy Scouts, or the NAACP. Glassdoor’s users are necessarily strangers to each other, because they are anonymous. Users do not so much “discuss” employment conditions as independently post their individual views. Although employers can comment on reviews, employees cannot comment directly on one another’s posts or otherwise engage in dialogue. The reviews of the subject company itself show no evidence that Glassdoor users share a common point of view, are engaged in a “common endeavor,” or have a unifying “mission.” Some reviewers are very positive, leaving comments like “This is a great company!” and “Unlimited opportunities for growth.” Others are negative: “Poor management;” “Nothing but a circus with total clowns;” “It was . . . very hard to watch the slow process of [the subject] cause delays of care daily.” 3

3 The First Amendment “does not require that every member of a group agree on every issue in order for the group’s policy to be ‘expressive association.’” Boy Scouts, 530 U.S. at 655. But this principle presupposes that a group has an official policy, as the petitioners in Boy Scouts did regarding acceptance of homosexual scout masters. See id. at 651–53. Here, neither Glassdoor nor its users have


In short, we are skeptical of the claim that Glassdoor’s users constitute an expressive association. But we certainly recognize enforcing the subpoena implicates their First Amendment rights, because they have a limited right to speak anonymously.

“[A]n author’s decision to remain anonymous . . . is an aspect of the freedom of speech protected by the First Amendment.” In re Anonymous Online Speakers, 661 F.3d 1168 , 1173 (9th Cir. 2011) (quoting McIntyre v. Ohio Elections Comm’n, 514 U.S. 334 , 342 (1995)). This is true whether the speech occurs online or offline. See id. However, the right to speak anonymously is not unlimited. Id. The degree of scrutiny applied to alleged infringements of the right to speak “varies depending on the circumstances and the type of speech at issue.” Id.

Here, the government seeks to unmask anonymous speakers in order to identify potential percipient witnesses in aid of a federal grand jury investigation into possible fraud. The speakers whose identities the government seeks may well be witnesses to this criminal activity, perhaps even participants in it. Presumably, once identified, federal agents will attempt to contact them to see whether they have any admissible evidence to offer in aid of the investigation and to serve those who do with a grand jury subpoena ad testificandum.


At virtually the same time, the Supreme Court and our court decided cases in which witnesses challenged grand

an official stance on the subject, nor can they—the site is designed to allow individual users to express their own opinions.


jury subpoenas that would have required them to identify individuals who wished to remain anonymous, but who might have had information about criminal activity or organizational membership.

In Branzburg v. Hayes, the Supreme Court held that a reporter—even one who has promised his sources anonymity—must cooperate with a grand jury investigation unless there is evidence that the investigation is being conducted in bad faith. 408 U.S. at 689–91, 707–08. The next day, in Bursey v. United States, we held that when a grand jury investigation into the activities of a group thought to be subversive “collides with First Amendment rights,” the government must satisfy a three-prong “compelling interest” test. 466 F.2d at 1083.

Glassdoor argues that we should apply Bursey’s compelling-interest test, while the government contends that Branzburg’s good-faith test controls. We hold that Branzburg supplies the correct test, for the reasons set forth below.


In Branzburg, the Supreme Court held that requiring newspaper reporters to appear and testify before grand juries regarding the identity of their sources did not violate the reporters’ First Amendment rights. 408 U.S. at 667. Branzburg was a consolidation of three cases involving journalists who had reported on illicit drug activity or the Black Panthers. Id. at 667, 672, 675. Each had been subpoenaed to testify before a grand jury, and each had refused to answer questions about the identity of his sources and, in some cases, what he had seen while reporting. Id. at 668, 673, 676.


The reporters argued that requiring them to identify their sources or testify to what they were permitted to see would deter sources from speaking to them in the future, hampering their ability to gather news. See id. at 679–80. The Court rejected this argument, declining to create for reporters “a testimonial privilege that other citizens do not enjoy.” Id. at 690. However, the Court stated that reporters would be entitled to relief from the courts if a grand jury investigation were “instituted or conducted other than in good faith.” Id. at 707–08.


Glassdoor insists that because it is not a reporter and is not claiming a “newsman’s privilege” for itself, Branzburg is inapplicable. It emphasizes that the Branzburg reporters sought to vindicate their own First Amendment right to gather news, while Glassdoor is asserting its users’ First Amendment right to speak and associate anonymously. But this is a distinction without a difference. At its core, Glassdoor’s argument is very similar to the one rejected in Branzburg.

In Branzburg, the reporters argued that “to gather news it is often necessary to agree . . . not to identify the source of information,” and that if reporters were forced to identify anonymous sources, they and other potential sources would “be measurably deterred from furnishing publishable information, all to the detriment of the free flow of information protected by the First Amendment.” Id. at 679– 80.

Although Glassdoor is not in the news business, as part of its business model it does gather and publish information from sources it has agreed not to identify. It argues that “[a]nonymity is an essential feature of the Glassdoor


community,” and that “if employees cannot speak anonymously, they often will not speak at all,” which will reduce the availability of “information about what it is like to work at a particular job and how workers are paid.” In other words, forcing Glassdoor to comply with the grand jury’s subpoena duces tecum will chill First Amendment- protected activity. This is fundamentally the same argument the Supreme Court rejected in Branzburg.

Furthermore, Branzburg makes it clear that Glassdoor’s users do not have a First Amendment right not to testify before the investigating grand jury about the comments they initially made under the cloak of anticipated anonymity. See id. at 695 (“[I]f the authorities independently identify the informant, neither his own reluctance to testify nor the objection of the newsman would shield him from grand jury inquiry . . . .”). Therefore, Glassdoor cannot refuse to turn over its users’ identifying information on the grounds that it is protecting its users’ underlying rights.


Glassdoor also argues that Branzburg is limited to “news gathering,” and that because Glassdoor is asserting its users’ rights to speak and associate anonymously, Branzburg is inapplicable. It’s true that the Branzburg Court stated that “[t]he sole issue” before it was “the obligation of reporters to respond to grand jury subpoenas.” Id. at 682. But the Court’s full statement was that “[t]he sole issue before” it was “the obligation of reporters to respond to grand jury subpoenas as other citizens do and to answer questions relevant to an investigation into the commission of crime.” Id. (emphasis added). The Court went on to say that “[c]itizens generally are not constitutionally immune from grand jury subpoenas; and neither the First Amendment nor any other constitutional provision protects the average


citizen from disclosing to a grand jury information that he has received in confidence.” Id.

In rejecting the reporters’ arguments for creating a constitutional “reporter’s privilege,” the Court stated that “the only testimonial privilege . . . that is rooted in the Federal Constitution is the Fifth Amendment privilege against compelled self-incrimination.” Id. at 689–90 (emphasis added). It expressly declined to “create another by interpreting the First Amendment to grant newsmen a testimonial privilege that other citizens do not enjoy.” Id. at 690.

Thus, although the Branzburg Court was responding to a newsgathering claim, it made clear that the First Amendment does not provide reporters—or anyone else—with a privilege against responding to a grand jury’s inquiries. Therefore, it doesn’t matter whether the underlying claim is related to newsgathering, speech, or association. These are all First Amendment-protected activities, but none of them will prevent an individual from being required to cooperate with a good-faith grand jury investigation. Only if a witness has a legitimate claim of self-incrimination under the Fifth Amendment may he refuse to answer questions or supply information relevant to the investigation.


One factor does distinguish Glassdoor’s users from the anonymous sources in Branzburg: there, at least two of the reporters had promised not to reveal the identities of those they observed while reporting. Id. at 667–68, 672–73. Thus, their sources reasonably expected anonymity.

But here, Glassdoor’s Privacy Policy puts its users on notice before their first submission is posted that their


identifying information could be revealed to the government in response to a subpoena or court order. The fact that Glassdoor’s users do not have a reasonable expectation of complete privacy further undermines Glassdoor’s contention that enforcing the subpoena would violate its users’ rights to anonymous speech or association.


Bursey is of no help here. There, a district court had held two staffers of the Black Panther newspaper in contempt for refusing to answer certain questions from a grand jury that was investigating a death threat against President Richard Nixon. Bursey, 466 F.2d at 1065. The questions propounded to the witnesses during their grand jury appearances went well beyond any legitimate connection to federal crimes.

We reversed, stating that “[w]hen governmental activity collides with First Amendment rights” in the context of a grand jury investigation, potential witnesses may not be compelled to answer questions unless the government establishes: 1) that the government has an “immediate, substantial, and subordinating” interest in the subject matter of the investigation; 2) “that there is a ‘substantial connection’ between the information it seeks” and its compelling interest in the investigation; and 3) that “the means of obtaining the information is not more drastic than necessary” to advance the government’s interest. Id. at 1083.

The Bursey court recognized the government’s compelling interest in investigating threats against the President, but held that the grand jury’s questions regarding the inner workings of the Black Panther newspaper and the identities of staff members who published the paper lacked


a “substantial connection to the compelling subject matter of the investigation.” Id. at 1086–88. Therefore, the witnesses could not be compelled to answer the objectionable questions. Id. at 1088.

The government petitioned for rehearing en banc, arguing that the court’s reasoning in Bursey was inconsistent with the Supreme Court’s recent decision in Branzburg. Id. at 1090. We rejected that argument for two reasons. First, “[t]he press function with which the [Branzburg] Court was concerned was news gathering,” but “[n]ews gathering [was] not involved” in Bursey. Id. Second, Bursey’s holding that the government must demonstrate a “‘substantial connection’ between the information sought and the criminal conduct which the Government was investigating” was not inconsistent with Branzburg. Id. at 1090–91. Thus, the Bursey panel ruled that its decision was in harmony with Branzburg. Id. at 1091.


Glassdoor nonetheless argues that Bursey controls and compels us to quash the subpoena. We are unpersuaded. The circumstances that led to our decision in Bursey are very different than those here. In Bursey, there was evidence that the government was engaged in a fishing expedition designed to gather as much background information as possible about the activities of a dissident group. A legitimate grand jury investigation that had been launched in response to a threat against the President had turned into “a general probe of the affairs of the Black Panther Party.” Id. at 1065. The grand jury had questioned Black Panther staffers about the “internal management of the paper,” demanding that they name “the people who edit the paper,” “the people who normally work on the newspaper,” “the person who is the current editor-in-chief of the newspaper,”


“the person currently responsible for distributing the newspaper,” and “the people . . . who do layout work for the paper.” Id. at 1066, 1068–69. They also asked the staffers to identify Black Panthers whose photographs had appeared in a different magazine. Id. at 1065–66.

Here, on the other hand, there is no evidence that the government is on an improper fishing expedition. Glassdoor is not being asked about the inner workings of Glassdoor itself, or being required to identify users simply because they have Glassdoor accounts. To the contrary, the government has only requested the identifying information of users who appear to have relevant information about the manner in which the subject of the grand jury’s investigation administered its government contracts. And it has narrowed its request for information from all 125 users who posted reviews about the subject to the eight whose reviews arguably describe what may be fraudulent behavior.


Since Bursey was decided, we have consistently applied Branzburg’s good-faith test rather than Bursey’s compelling-interest test when deciding First Amendment challenges to good-faith grand jury inquiries. Glassdoor acknowledged in its briefs and confirmed at oral argument that it is not alleging any bad faith on the part of the Arizona grand jury conducting this particular investigation.

In Lewis v. United States, we applied Branzburg to affirm a judgment of contempt against a radio station manager who refused to comply with a grand jury subpoena. 501 F.2d 418 , 422–23 (9th Cir. 1974) (Lewis I). Lewis I did not mention Bursey. See id. The following year, we affirmed a different contempt judgment against the same appellant, citing Branzburg and stating that reporters would


be protected from grand jury subpoenas only “where a grand jury investigation is instituted or conducted other than in good faith.” Lewis v. United States, 517 F.2d 236 , 237–38 (9th Cir. 1975) (Lewis II). Again, there was no mention of Bursey.

In In re Grand Jury Proceedings (Scarce), a doctoral student claimed that a “‘scholar’s privilege’ . . . akin to that of a reporter” should protect him from being forced to answer questions from a grand jury investigating criminal activity at a research facility. 5 F.3d 397 , 398 (9th Cir. 1993). We disagreed. Relying on Branzburg, we held that because there was no evidence of bad faith, the appellant was required to cooperate with the grand jury. Id. at 400.

Since Bursey was decided in 1972, we have never relied on it to quash a grand jury subpoena such as the one before us, and we see no reason to do so here.


Glassdoor argues that our denial of the government’s petition for rehearing in Bursey fatally undermines the government’s contention that Branzburg applies here. According to Glassdoor, in Bursey, we “considered and rejected the argument that Branzburg overrules Bursey.” Therefore, Bursey controls and the district court should have applied its three-prong, compelling-interest test. However, the government does not argue, and we do not hold, that Branzburg overruled Bursey, only that Branzburg should be applied here.

We reached a similar conclusion in Scarce, where we determined that Bursey does not preclude application of Branzburg’s good-faith test. See 5 F.3d at 402. In Scarce, we recognized that balancing the government’s interests


against an individual’s First Amendment rights was appropriate in Bursey, where there was not a substantial connection between the information the government sought and its investigation. Id. But because there was a substantial connection in Scarce, Bursey was inapplicable and no balancing of interests was necessary. See id.

Similarly, here, there is a substantial connection between the information the government seeks and the criminal conduct it is investigating. Therefore, Branzburg applies. Neither Bursey’s original holding nor the panel’s subsequent denial of the government’s petition for rehearing compels a different result.


Branzburg’s good-faith test is also more appropriate than Bursey’s compelling-interest test given the nature and importance of grand jury proceedings. Grand juries, which are responsible for “determining whether or not a crime has been committed,” occupy “a unique role in our criminal justice system.” United States v. Real Enters., Inc., 498 U.S. 292 , 297 (1991). “[T]he law presumes, absent a strong showing to the contrary, that a grand jury acts within the legitimate scope of its authority.” Id. at 300. “A grand jury subpoena issued through normal channels is presumed to be reasonable, and the burden of showing unreasonableness” is “on the recipient who seeks to avoid compliance.” Id. at 301.

In Real Enterprises, the Supreme Court held that the relevancy, admissibility, and specificity standards that apply to trial subpoenas do not apply to grand jury subpoenas, in part because “[a]ny holding that would saddle a grand jury with minitrials and preliminary showings would . . . impede its investigation and frustrate the public’s interest in the fair and expeditious administration of the criminal laws.” Id. at


298–99 (quotation omitted). The Court also found that “[r]equiring the Government to explain in too much detail the particular reasons underlying a subpoena threatens to compromise ‘the indispensable secrecy of grand jury proceedings.’” Id. at 299 (quoting United States v. Johnson, 319 U.S. 503 , 513 (1943)).

Requiring the government to satisfy the three prongs of Bursey’s compelling-interest test every time a grand jury subpoena implicates the right to anonymous speech or association would clearly invite the kind of “minitrials and preliminary showings” the Supreme Court rejected in Real Enterprises. And requiring the government to disclose the specifics of a grand jury investigation to establish that a subpoena did not violate the First Amendment would threaten “the indispensable secrecy of grand jury proceedings.” See id.

Branzburg’s good-faith test does not pose the same dangers.


Because we hold that Branzburg controls, the only question is whether there is evidence that the grand jury is acting in bad faith. The Branzburg majority noted that grand jury investigations “instituted or conducted other than in good faith” pose “wholly different” First Amendment issues than the good-faith investigations at issue there and stated that “[o]fficial harassment of the press . . . would have no justification.” 408 U.S. at 707–08.

Justice Powell elaborated in a concurrence, explaining that a reporter “will have access to the court on a motion to quash” if he “is called upon to give information bearing only a remote and tenuous relationship to the subject of the


investigation, or if he has some other reason to believe that his testimony implicates confidential source relationships without a legitimate need of law enforcement.” Id. at 710 (Powell, J., concurring).

We incorporated all those aspects of bad faith into our analysis in Scarce, holding that the appellant there was “not entitled to a First Amendment privilege” because he did not argue that the grand jury’s questions “were posed in bad faith, that they had a tenuous relationship to the subject of the investigation, that law enforcement did not have a legitimate need for the information, or that they were posed as a means of harassment.” 5 F.3d at 400.

Here, Glassdoor does not assert that the grand jury investigation is being conducted in bad faith, or that the subpoena is intended to harass. It does contend that there is only a tenuous connection between the information the grand jury seeks and the subject of its investigation, arguing that “[t]here is no substantial connection between the identities of employees the government seeks to obtain and its investigation into criminal fraud.” This assertion is belied by the record.

The grand jury is investigating the subject government contractor for fraud, waste, and abuse of federal funds. Each of the employees whose contact information the government seeks posted a review of the subject that referenced potentially fraudulent conduct. Some postings are very specific as to the nature of the misconduct. For example, one reviewer wrote, “Highly inefficient system that focuses more on call quotas ([the subject] makes revenue by charging the VA for the calls that are made). There is no intention of designing a more efficient system to assist Veterans because that would reduce the number of calls made . . . which equals less revenue.” Others are less specific: “Everything is


supposedly ‘for the Veterans’ but all they care about is numbers.”

Read in isolation, the less specific reviews might not indicate that the reviewers have information about fraud. But when read together with the other reviews—several of which describe a scheme whereby the subject seeks to maximize call volume without delivering improved service to veterans—the statement that “all they care about is numbers” supports an inference that the reviewer has observed fraudulent conduct.

The information the government seeks will allow the grand jury to contact (presumably through interviews by federal agents sworn to assist the grand jury in its inquiries) and then question employees who have observed potentially fraudulent behavior by the company. Thus, there is a clear connection between the nature of the investigation—waste, fraud, and abuse by the subject—and the information the government seeks—the identity of potential witnesses to that fraud and abuse.

Because Glassdoor has neither alleged nor established bad faith on the part of the government in its investigation, under Branzburg, enforcement of the subpoena duces tecum to identify potential witnesses in aid of its inquiries does not violate the First Amendment rights of Glassdoor’s users.


Glassdoor argues that if even if we choose not to reverse the district court’s denial of the motion to quash outright, we should remand with instructions for the district court to hold an evidentiary hearing. But Glassdoor has not shown that any other evidence is necessary to rule on its objection.


Absent a colorable allegation of bad faith on the part of the government, and without a credible argument that there is a tenuous relationship between the information Glassdoor holds and the focus of the investigation, under Branzburg, Glassdoor’s motion to quash is unavailing. But even if we were to apply the test in Bursey, we would still affirm the denial of the motion to quash on this record.

The government clearly has a legitimate and compelling interest in the grand jury’s investigation into possible violations of federal law. “[T]he investigation of crime by the grand jury implements a fundamental governmental role of securing the safety of the person and property of the citizen.” Branzburg, 408 U.S. at 700; see also Dole v. Serv. Emp. Union, AFL-CIO, Local 280, 950 F.2d 1456 , 1461 (9th Cir. 1991) (“[T]here is little doubt that the Department’s purpose of investigating possible criminal violations of [a federal law] serves a compelling governmental interest.”).

The grand jury is conducting a criminal investigation into alleged fraud and abuse by a government contractor that administers veterans’ healthcare programs. If the allegations are true, the subject is not only misusing taxpayer funds, it is deliberately making it more difficult for veterans to access healthcare. The government’s interest in investigating such behavior is self-evident.

The district court correctly ruled that there is a substantial connection between the subject matter of the investigation and the identifying information of the eight users whose Glassdoor posts allude to potentially fraudulent behavior. We agree. Any incidental infringement on Glassdoor’s users’ First Amendment rights is no more drastic than necessary to vindicate those compelling interests.


We see no reason to remand for further development of the record. Doing so would only delay the grand jury investigation even further for no good reason, and the Supreme Court has cautioned against presenting grand jury investigations with “procedural delays and detours” that “frustrate the public’s interest in the fair and expeditious administration of the criminal laws.” Real Enters., 498 U.S. at 298–99 (quoting United States v. Dionisio, 410 U.S. 1 , 17 (1973)).

“[T]he longstanding principle” that the public “has a right to every man’s evidence . . . is particularly applicable to grand jury proceedings.” Branzburg, 408 U.S. at 688 (internal quotations and citations omitted). This case presents no reason to carve out an exception to that principle.


Back to top