concurring in part and concurring in the judgment:
I write separately to explain my reasoning in this challenging case.
, FOIA allows for an award of attorney fees to eligible parties who can show their entitlement to fees. Where a FOIA plaintiff prevails on the merits of her claim, eligibility for fees is straightforward. Where a FOIA plaintiff does not prevail on the merits of her claim, but the Government nonetheless releases some or all of the information sought through a unilateral disclosure, eligibility for fees is less straightforward. The leading test for analyzing eligibility in these circumstances asks whether the plaintiffs litigation was a “substantial causative effect” of the disclosure of the information. See Church of Scientology of Cal. v. U.S. Postal Serv., 700 F.2d 486, 489 (9th Cir. 1983). In keeping with its name, the substantial causative effect test has long incorporated a causation requirement. See, e.g., Long v. IRS, 932 F.2d 1309, 1313 (9th Cir. 1991). We sometimes refer to this as a “catalyst theory” of recovery. Oregon Nat. Desert Ass’n v. Locke, 572 F.3d 610, 612, 615 (9th Cir. 2009).
*887This ease concerns whether First Amendment Coalition, a FOIA plaintiff, is eligible for an award of attorney fees. The district court concluded that First Amendment Coalition did not prevail on the merits or satisfy the “substantial causative effect” test for a catalyst recovery, and was therefore ineligible for an award of fees. The appeal of that decision presents three distinct questions: (1) whether the catalyst theory and its “substantial causative effect” test still includes a causation requirement under the amended text of FOIA; (2) whether the district court clearly erred in its factual finding on causation; and (3) whether First Amendment Coalition is eligible for an award of attorney fees.
On the first question, I join in the analysis of part II of the opinion, finding that recovery under a catalyst theory continues to require causation. On the second question, I do not join in the opinion’s analysis, because I do not agree that the district court clearly erred in its causation finding. On the third question, I conclude that First Amendment Coalition is eligible for fees. The Department of Justice (the DOJ) prevented First Amendment Coalition from prevailing on the merits, and the district court erred by failing to take into account the DOJ’s conduct when analyzing eligibility. I therefore concur in reversing the district court’s judgment on' First Amendment Coalition’s eligibility for a fee award, and in remanding to consider- First Amendment Coalition’s entitlement to fees,
I. Background
In 2010, lawyers in the Office of Legal Counsel (OLC) in the DOJ wrote two memoranda analyzing the targeted killing of an American citizen abroad: a memorandum from the OLC to the U.S. Department of Defense '(the OLC-DOD Memorandum) and a memorandum from the OLC to the Central Intelligence Agency (the OLC-CIA Memorandum). These memoranda-were not public, and various Government agencies refused to confirm or deny their existence. When First Amendment Coalition submitted its FOIA request in 2011, these two memoranda were within the scope of the request. DOJ did not provide these memoranda to First Amendment Coalition, or confirm or deny their existence. First Amendment Coalition then litigated in pursuit of the still-undisclosed and unacknowledged responsive materials in the United States District Court for the Northern District of California (NDCA). Other parties brought similar FOIA actions in the United States District Court for the Southern District of New York (SDNY). /. .. '
The SDNY plaintiffs lost on the merits of their suit in early 2013. N.Y. Times Co. v. U.S. Dep’t of Justice (NY Times I), 915 F.Supp.2d 508, 516-18 (S.D.N.Y. 2013). Later that year, the United States Government acknowledged the OLC-DOD Memorandum after a series of other related disclosures, most notably an internal “White Paper” summarizing the OLC legal rationales. The DOJ and other agencies, however, did not make the OLC-DOD Memorandum public, and did not disclose the existence of the OLC-CIA Memorandum.
In the NDCA litigation, the DOJ and First Amendment Coalition filed cross-motions for summary judgment regarding the release of the materials First Amendment Coalition had requested. .The DOJ did not mention the White Paper in the opening brief for its second motion -for summary judgment. After First Amendment Coalition emphasized the leaked White Paper -in its briefing, the DOJ in its reply to its motion for summary judgment and opposition to First Amendment Coalition’s cross-motion stated: “The White Paper is a draft document that was originally provided in *888confidence to Congress; the Executive Branch later acknowledged the draft document.” The DOJ did not inform First Amendment Coalition or the district court that it had provided the White Paper to a news organization called Truthout in response to Truthout’s separate FOIA request. See N.Y. Times Co. v. U.S. Dep’t of Justice (NY Times III), 756 F.3d 100, 110 n.9 (2d Cir. 2014) (noting the United States also never informed the ACLU, a plaintiff in the SDNY litigation of this release). This action constituted an intentional public release and disclosure of the White Paper, rather than a mere acknowledgment, and that distinction proved critical.
In early 2014, the NDCA district court ruled against First Amendment Coalition on the merits of the cross-motions for summary judgment. In its order, the district court observed that First Amendment Coalition “makes much of the fact that the unclassified White Paper prepared for Congress has been leaked and acknowledged by the Government. However, there has been no ‘official disclosure’ of the White Paper.” The district court’s statement that there was no official disclosure was incorrect given the prior DOJ release of the White Paper in response to the Truthout FOIA request. Id.
Ten days after the NDCA district court ruled against First Amendment Coalition on the merits, the Second Circuit reversed the SDNY district court and ordered the release of a redacted version of the OLC-DOD Memorandum. N.Y. Times Co. v. U.S. Dep’t of Justice (NY Times II), 752 F.3d 123, 126 (2d Cir.), opinion revised and superseded, 756 F.3d 100 (2d Cir. 2014). In its opinion, the Second Circuit determined that the OLC-DOD Memorandum’s legal analysis was no longer exempt from FOIA disclosure because of the release of the White Paper and public statements by United States officials. Id. at 141. The Second Circuit specifically noted that the White Paper made a dispositive difference, focusing on the effect of the White Paper in showing waiver of the United States’ privilege arguments:
Even if these statements assuring the public of the lawfulness of targeted killings are not themselves sufficiently detailed to establish waiver of the secrecy of the legal analysis in the OLC-DOD Memorandum, they establish the context in which the most revealing document, disclosed after the [SDNY] District Court’s decision, should be evaluated. That document is the DOJ White Paper[.]
Id. at 138; see id. at 141 (“Whatever protection the legal analysis might once have had has been lost by virtue of public statements of public officials at the highest levels and official disclosure of the DOJ White Paper.”).
The Second Circuit’s opinion also included the following paragraph, referencing the NDCA litigation:
The recent opinion of the District Court for the Northern District of California, ... denying an FOIA request for the OLC-DOD Memorandum, is readily distinguishable because the Court, being under the impression -that “there has been no ‘official disclosure’ of the White Paper,” .... did not assess its significance, whereas in our case, the Government has conceded that the White Paper, with its detailed analysis of legal reasoning, has in fact been officially disclosed.
Id. at 139 (internal citations omitted).
The Second Circuit subsequently published the OLC-DOD Memorandum in a revised decision. See NY Times III, 756 F.3d at 104. Shortly after, the DOJ disclosed the existence of and voluntarily released the OLC-CIA Memorandum to plaintiffs in the SDNY and to First *889Amendment Coalition. The precise reasons why the DOJ voluntarily disclosed the existence of and released the OLC-CIA Memorandum are in dispute between the parties and there is almost no direct evidence of why the DOJ made that decision.
II. Catalyst Theory
First Amendment Coalition argues that the NDCA litigation substantially caused the release of the OLC-DOD Memorandum and the OLC-CIA Memorandum. First Amendment Coalition’s principal theory for why it had a “substantial causative effect” on the DOJ’s release of both the OLC-DOD and OLC-CIA memoranda is that its NDCA litigation — and the prospect of Ninth Circuit review — affected the Government’s choices. First Amendment Coalition points to the DOJ’s decision not to contest the release of the OLC-DOD memorandum the Second Circuit ordered, and the DOJ’s decision to release the OLC-CIA memorandum before any federal court ordered the DOJ to do so. The NDCA district court rejected First Amendment Coalition’s arguments, finding that the DOJ “in this case released the documents largely as a result of the Second Circuit’s ruling in NY Times, not as a result of the ruling in this case.”1
The district court’s ruling on causation is a factual finding, and we review for clear error. See Ass’n of Cal. Water Agencies v. Evans, 386 F.3d 879, 886 (9th Cir. 2004). Here, there is evidence to support First Amendment Coalition’s position, but there is also evidence consistent with the district court’s finding.2 When one of two plausible explanations supports the district court finding, there cannot be a “definite and firm conviction that a mistake has been committed.” N. Queen Inc. v. Kinnear, 298 F.3d 1090, 1095 (9th Cir. 2002) (quoting Allen v. Iranon, 283 F.3d 1070, 1076 (9th Cir. 2002)). Therefore, in my view, we must accept the district court’s finding of causation, even if there is some cause to doubt it. Ass’n of Cal. Water Agencies, 386 F.3d at 886. Because I think we must uphold the district court’s factual finding with respect to both responsive memoran-da, I conclude First Amendment Coalition cannot show eligibility for fees under a catalyst theory. See Long, 932 F.2d at 1313. I therefore do not join in the opinion to the extent it reverses the district court while still applying a catalyst theory.
III. A Different Ground for Fee Eligibility
In the majority of FOIA fee cases, the catalyst theory of recovery will be the appropriate way to analyze the plaintiffs eligibility for fees. But, to my mind, a *890catalyst theory of recovery does not suit the facts of this case.
In my view, however, there is no need to rely on a catalyst theory or re-write the “substantial causative effect” test.3 We can instead resolve the case by returning to the purposes underlying the test, and re-framing them for the unique circumstances presented here. After doing so, it is clear that First Amendment Coalition has another viable theory for eligibility, equally consistent with FOIA’s text and purpose.
At the time of its adoption, the substantial causative effect test established “that a court judgment is not a prerequisite for an attorney fees award under” FOIA. Nationwide Bldg. Maint., Inc. v. Sampson, 559 F.2d 704, 708-09 (D.C. Cir. 1977). In practice, the test recognizes that plaintiffs in FOIA suits- do not need to secure a judgment on the merits to receive attorney fees. The purpose behind this rule is simple: when the Government takes unilateral action — completely outside the plaintiffs control' — such action should not -prevent the plaintiff from receiving compensation for a meritorious claim. See Church of Scientology, 700 F.2d at 492 (noting that “the mere fact that defendants have voluntarily released documents does not preclude an award of attorney’s fees to the plaintiff’).
To illustrate, suppose the Government, on the eve of defeat in FOIA litigation, unilaterally releases the requested information, thereby mooting the case and preventing the plaintiffs from securing a judgment. The' plaintiffs’ efforts in litigation compelled the disclosure of information, and the plaintiffs would have prevailed absent the Government’s conduct. Accordingly, we would award attorney fees to the plaintiffs under á catalyst theory, even though they did not secure a judgment on the merits. The instant case presents a different variation on this situation, but the same considerations should still apply.
First Amendment Coalition and the SDNY FOIA plaintiffs sought the same underlying materials for the same reasons. Thp DOJ then took a unilateral action that prevented .-First .Amendment Coalition from prevailing on the merits, by failing to describe the White Paper candidly to the NDCA district court, Because the DOJ did not inform First Amendment Coalition or the NDCA district court about the official release of the White Paper in response to the .FOIA request, the DOJ deprived the NDCA district court of the opportunity to ■ rule on. a complete record. The district court even admitted that it “may not have been fully apprised of. the facts surrounding the White Paper” in making its earlier ruling, a conclusion reached only with the clarity provided by the Second Circuit decision. The Second Circuit recognized the White Paper was critical,' and that the DOJ’s representations to- the NDCA district court had differed in a material way. Something has gone wrong when a district court learns of relevant information first from another court, let alone material, case-dispositive information. E.g., United States v. Shaffer Equip. Co., 11 F.3d 450, 459 (4th Cir. 1993) (holding that failure to disclose information that “could conceivably have affected the outcome of the litigation violates the general duty of candor that attorneys owe as officers of the court”).
First Amendment Coalition failed to prevail primarily because of unilateral Government actions, outside its control. The district court erred in not recognizing *891the significance of the DOJ’s litigation conduct for the FOIA eligibility analysis. By characterizing the White Paper inconsistently between two federal fora, the DOJ deprived First Amendment Coalition of the fair chance for the court to hear its arguments on the merits. The first and only federal court to consider the merits of whether the official disclosure of the White Paper affected the FOIA analysis concluded that the White Paper made a decisive difference. In the course of its decision, the Second Circuit also observed that the only apparent difference between the case before it and the case before the NDCA district court was that the NDCA district court had a different understanding of the White Paper’s significance. The district court here did not draw its understanding from the ether; it relied on the Government’s representations. If the NDCA district court had in front of it. the same information available to the Second Circuit, the NDCA district court likely would have reached the same result as the Second Circuit — or we would have, on appeal.
Under these circumstances, the same rationales apply as- in “substantial causative effect” cases. First Amendment Coalition acted in good faith, brought a timely action, and was correct oh the merits. A ruling in First Amendment Coalition’s favor would likely have made First Amendment Coalition eligible without even the need to rely on any alternate theory of recovery. See 5 U.S.C. § 552(a)(4)(E)®. First Amendment Coalition was unable to recover because of the DOJ’s unilateral action in how the DOJ characterized the White Paper to the court. The DOJ prevented First Amendment Coalition from prevailing and establishing its eligibility via a favorable judgment, id. § 552(a)(4)(E)®, and we therefore may look to alternate bases establishing First Amendment Coalition’s eligibility to recover, id. § 552(a)(4)(E)(ii). The district court erred as a matter of law by not accounting for the DOJ’s actions when analyzing First Amendment Coalition’s eligibility for recovery, and by limiting its analysis to actual causation.
Accordingly, I would reverse the district court for legal error, not for its underlying causation finding. When we take into account the DOJ’s conduct, First Amendment Coalition has established its eligibility for an award of fees. I therefore join my colleagues in reversing the district court ruling on eligibility, and remanding for a determination of First Amendment' Coalition’s entitlement to fees.4
. It appears the district court was referring to NY Times II.
. For instance, we have two available explanations for why the DOJ ultimately voluntarily disclosed the OLC-CIA Memorandum to First Amendment Coalition. The first explanation is that the DOJ concluded that the waiver determinations the Second Circuit made regarding the OLC-DOD Memorandum applied with equal force to the OLC-CIA Memorandum. See NY Times III, 756 F.3d at 117. In other words, in response to the Second Circuit decision and subsequent SDNY district court order, the DOJ conceded. The second explanation is that the DOJ wanted to end the NDCA litigation. (I note that the DOJ had not yet responded to the NDCA district court’s order to provide an explanation for why the DOJ conceded official disclosure of the White Paper in one federal forum but made no mention of official disclosure in another. The district court had ordered the DOJ to explain this inconsistency if the case was not moot, and disclosing the OLC-CIA Memorandum prevented the DOJ from needing to respond.) The first explanation supports the district court’s finding of causation; the second explanation suggests the district court’s finding of causation was in error, and supports First Amendment Coalition’s eligibility for fees.
. As noted above, I join in the opinion's conclusion that causation remains a requirement of the "substantial causative effect" test for a catalyst theory of recovery. See Opinion, Part II.
. In analyzing whether First Amendment Coalition is entitled to fees, as opposed to merely eligible, the district court will be able to consider agency behavior or reasonableness, See Button v. IRS, 718 F.3d 522, 527 (5th Cir. 2013); Miller v. U.S. Dep’t of State, 779 F.2d 1378, 1390 (8th Cir. 1985).