KAREN LECRAFT HENDERSON, Circuit Judge, dissenting: Over the past 15 years, we have remanded this case four times. During the same period, we have reversed the same district court twice in a nearly identical Freedom of Information Act (FOIA) case. That makes six opinions from this court. I share the majority’s displeasure with the resulting waste of judicial resources, especially because “fee litigation [is] one of the last things lawyers and judges should be spending their time on.” Baylor v. Mitchell Rubenstein & Assocs., P.C., 857 F.3d 939, 960 (D.C. Cir. 2017) (Henderson, J., concurring). Jefferson Morley, however, is not to blame for this “staggering” saga. Maj. Op. 2. But for the district court’s repeated misapplication of FOIA precedent, this case could have ended as early as 2006. If it had been correctly decided the first time, “Morley would already have his fees, and this litigation would have long since concluded.” Morley v. CIA, 719 F.3d 689, 693 (D.C. Cir. 2013) (Kavanaugh, J., concurring). Unfortunately, the district court got it wrong again. The majority, it appears to me, overlooks the district court’s latest errors in order to “bring the case to an end.” Maj. Op. 2. In the process, it distorts our settled fourfactor test for awarding attorney fees under FOIA and replaces it with a single-factor reasonableness inquiry of its own design. What’s worse, the majority misapplies its own test. It holds that the Central Intelligence Agency (CIA) reasonably declined to produce any documents in response to Morley’s FOIA request and instead directed him to another agency. The holding is plainly contrary to Tax Analysts v. DOJ, which declared that “an agency must itself make disclosable agency records available to the public and may not on grounds of administrative convenience avoid this statutory duty by pointing to another public source for the information.” 845 F.2d 1060, 1067 (D.C. Cir. 1988) (emphasis in original), aff’d, 492 U.S. 136 (1989). To avoid this precedent—and to explain away the district court’s contrary conclusions—the majority leans heavily on the standard of review, declaring that it requires “[d]eference piled on deference.” Maj. Op. 11. In my 2 view, my colleagues pile their deference far too high. Our abuse-of-discretion review, although forgiving, is not an empty formality: here, the district court’s discretion was constrained by our earlier opinions in this very case and by our closely related decision in Davy v. CIA, 550 F.3d 1155 (D.C. Cir. 2008). Because the district court failed to follow precedent and because it misapplied our four-factor test—for the third time— I believe it abused its discretion. Accordingly, I respectfully dissent. I. BACKGROUND In brief, the facts of this case are as follows: Morley is a journalist and news editor. On July 4, 2003, Morley submitted a request under FOIA to the CIA for “all records pertaining to CIA operations officer George Efythron Joannides.” The letter makes clear that Morley sought information connected to President John F. Kennedy’s assassination. The CIA responded in the beginning of November, 2003, with a letter explaining that the National Archives and Records Administration (“NARA”) had a public collection of CIA records related to the JFK assassination, which was searchable online. The CIA directed him to submit his request to NARA and did not release any records directly to Morley at that time. Morley subsequently filed suit in this Court on December 16, 2003, to enforce his FOIA request. After further processing of the request, along with an appeal up to our Circuit, the CIA ultimately provided Morley with a total of 524 responsive records (some of which were 3 segmented and/or redacted). Of those records, 113 were from the files the CIA previously had transferred to NARA. Morley v. CIA, 245 F. Supp. 3d 74, 76 (D.D.C. 2017) (Morley X) (quoting Morley v. CIA (Morley VIII), 59 F. Supp. 3d 151, 153-54 (D.D.C. 2014)). The majority truncates the history of this case, which, with this appeal, marks Morley XI. I believe more detail is needed to explain how our earlier decisions should have limited the district court’s discretion here. In response to Morley’s initial FOIA request, the CIA referred him to NARA without producing any of the requested documents. Morley filed suit. After the CIA produced three documents in full and 112 documents in segregable form, the district court granted its motion for summary judgment. Morley v. CIA (Morley I), 453 F. Supp. 2d 137, 142 (D.D.C. 2006). We affirmed in part but reversed in the main, giving seven remand instructions. Morley v. CIA (Morley II), 508 F.3d 1108, 1129 (D.C. Cir. 2007). We instructed the district court to direct the CIA to: (1) search operational files; (2) search records transferred to NARA; (3) supplement its explanation regarding missing monthly reports; (4) provide details regarding the scope of its search; (5) explain why withheld information was not segregable; (6) substantiate its Glomar response; and (7) provide further justification for its reliance on FOIA Exemptions 2, 5 and 6. See id. As most relevant here, we explained that FOIA reflects “a ‘settled policy’ of ‘full agency disclosure,’” id. at 1119 (quoting Tax Analysts, 845 F.2d at 1064), and “an agency has ‘withheld’ a document under its control when, in denying an otherwise valid request, it directs the requester to a place outside of the agency where the document may be publicly available,” id. 4 (alterations omitted) (quoting DOJ v. Tax Analysts, 492 U.S. 136, 150 (1989)). Because the CIA directed Morley to NARA rather than searching its own records, we held that it had failed to meet its duties under FOIA. Id. at 1120. Over the course of two years on remand, the CIA released 409 additional documents to Morley. The district court then granted the CIA’s renewed motion for summary judgment. Morley v. CIA (Morley III), 699 F. Supp. 2d 244 (D.D.C. 2010). On appeal, we affirmed in large part but remanded the case so the district court could examine Exemption 21 in light of a thenrecently decided Supreme Court case, Milner v. Department of Navy, 562 U.S. 562 (2011). Morley v. CIA (Morley V), 466 F. App’x 1 (D.C. Cir. 2012) (per curiam). On remand, the district court dismissed the case as moot. Morley v. CIA (Morley VI), No. 03-2545, 2013 WL 140245 (D.D.C. Jan. 9, 2013). Thus ended Morley’s merits dispute. While the CIA was defending on the merits of Morley’s FOIA suit, the same district judge heard an attorney’s fees dispute in another case involving a journalist (William Davy) who sought documents from the CIA regarding President Kennedy’s assassination. Davy v. CIA (Davy I), 357 F. Supp. 2d 76 (D.D.C. 2004). After obtaining documents through a consent order, Davy sought attorney’s fees. Davy v. CIA (Davy II), 456 F.3d 162 (D.C. Cir. 2006). The district court first denied Davy’s request. Id. at 163. We reversed, concluding that Davy was a prevailing party. Id. at 166. We then remanded for the district court to determine whether Davy was entitled to fees and, if so, to calculate those fees. Id. 1 Exemption 2 protects from disclosure agency material that is “related solely to the internal personnel rules and practices of an agency.” 5 U.S.C. § 552(b)(2). 5 On remand, the district court again denied Davy’s request for fees. Davy v. CIA (Davy III), 496 F. Supp. 2d 36 (D.D.C. 2007). Eventually, applying our “familiar four-factor test,” we again reversed. Davy v. CIA (Davy IV), 550 F.3d 1155, 1157 (D.C. Cir. 2008).2 We first addressed the “public benefit” factor, noting that the Kennedy assassination was an “event of national importance” and the information Davy obtained might help the public make “vital political choices.” Id. at 1160. We then examined the second and third factors in tandem and concluded that, although Davy may enjoy some pecuniary benefit from publishing books or articles as a result of his search, “that alone cannot be sufficient to preclude an award of attorney’s fees under FOIA.” Id. at 1160. Therefore, we held, the district court abused its discretion in finding that the second and third factors weighed against Davy. Id. at 1162. As to the fourth factor, we explained that, because the CIA failed even to respond to Davy’s request for documents until after he filed suit, the CIA was unreasonable in its initial withholding. Id. at 1163. Accordingly, all four factors favoring Davy, we remanded to the district court for it to award fees. Thus ended Davy’s fees dispute. Meanwhile, Morley filed an application for attorney’s fees. Morley v. CIA (Morley IV), 828 F. Supp. 2d 257 (D.D.C. 2011). The district court denied his request, finding that all 2 As discussed infra, the four factors are: (1) the public benefit derived from the case; (2) the commercial benefit to the plaintiff; (3) the nature of plaintiff’s interest in the records; and (4) the reasonableness of the agency’s withholding of the requested documents. Davy IV, 550 F.3d at 1159. Some of our sister circuits have adopted the same four-factor test. See, e.g., Pietrangelo v. U.S. Army, 568 F.3d 341, 343 (2d Cir. 2009); Texas v. ICC, 935 F.2d 728, 730 (5th Cir. 1991); Church of Scientology of Cal. v. U.S. Postal Serv., 700 F.2d 486, 492 (9th Cir. 1983). 6 four factors weighed against him. Id. at 260. Continuing a trend, we vacated the decision and remanded. Morley v. CIA (Morley VII), 719 F.3d 689 (D.C. Cir. 2013) (per curiam). In doing so, we directed the district court to our previous opinion in Davy IV, which stated that records “about individuals allegedly involved in President Kennedy’s assassination serve a public benefit.” Id. at 690 (alterations omitted) (quoting Davy IV, 550 F.3d at 1159). Moreover, we quoted Davy IV’s instruction that the public-benefit factor should not “disqualify plaintiffs who obtain information that, while arguably not of immediate public interest, nevertheless enables further research ultimately of great value and interest, such as here the public understanding of a [p]residential assassination.” Id. (quoting Davy IV, 550 F.3d at 1162 n.3). We remanded for the district court to consider Davy IV but did not express any position on whether it should award fees. Id. On remand, the district court again denied fees. Morley VIII, 59 F. Supp. 3d at 153. After analyzing the 524 documents the CIA ultimately produced in response to Morley’s request, the district court held that the “litigation has benefited the public only slightly, if at all.” Id. at 158. Without providing detail, it decided that its “analysis of the other factors remain[ed] the same” and denied Morley’s application. Id. For the fourth time, we remanded, concluding that “the district court improperly analyzed the public-benefit factor by assessing the public value of the information received rather than the ‘potential public value of the information sought.’” Morley v. CIA (Morley IX), 810 F.3d 841, 842 (D.C. Cir. 2016) (quoting Davy IV, 550 F.3d at 1159). We declared that, when evaluated ex ante, “Morley’s request had potential public value.” Id. at 844. We remanded for the district court to evaluate all four factors anew. Id. at 845. 7 This protracted history brings us to the district court’s most recent fees order. Morley v. CIA, 245 F. Supp. 3d 74 (D.D.C. 2017) (Morley X). In view of Morley IX, the district court found that the “expectation-adjusted value of the public benefit that plaintiff sought to provide was small.” Id. at 77. The court saw the second and third factors as a “close call.” Id. at 78. Specifically, it found that Morley received “some compensation for writing news articles” and saved time and energy by not having to seek documents in NARA’s Kennedy collection. Id. at 77. “Thankfully,” the court concluded, “the final factor breaks the tie—it weighs heavily against Morley and is ultimately dispositive.” Id. at 78. Accordingly, the district court denied Morley’s motion for attorney’s fees a third time. Id. II. DISCUSSION Under the traditional “American Rule,” each party to a lawsuit pays its own attorney’s fees. See Alyeska Pipeline Serv. Co. v. Wilderness Soc’y, 421 U.S. 240, 245 (1975). FOIA creates a statutory exception to the American Rule; it provides that the “court may assess against the United States reasonable attorney fees . . . in any case under this section in which the complainant has substantially prevailed.” 5 U.S.C. § 552(a)(4)(E)(i). We have explained that “FOIA’s attorney’s fees provision . . . was designed to lower the ‘often . . . insurmountable barriers presented by court costs and attorney fees to the average person requesting information under the FOIA.’” Tax Analysts v. DOJ, 965 F.2d 1092, 1095 (D.C. Cir. 1992) (quoting Cuneo v. Rumsfeld, 553 F.2d 1360, 1364 (D.C. Cir. 1977)). “[T]he award of FOIA counsel fees has as its fundamental purpose the facilitation of citizen access to the courts, and should not be subject to a grudging application.” First Amendment Coal. v. DOJ, 878 F.3d 1119, 1130 (9th Cir. 2017) (internal quotation marks omitted). 8 We review the district court’s application of the four-factor test for abuse of discretion. Id. at 1158. The district court’s discretion has two important limits. First, it is constrained by precedent. Koon v. United States, 518 U.S. 81, 100 (1996) (“A district court by definition abuses its discretion when it makes an error of law.”). Second, the district court’s discretion is limited by the mandate rule, which provides that “an inferior court has no power or authority to deviate from the mandate issued by an appellate court.” Briggs v. Penn. R.R., 334 U.S. 304, 306 (1948); see United States v. Kpodi, 888 F.3d 486, 491 (D.C. Cir. 2018) (“A district court commits legal error and therefore abuses its discretion when it fails to abide by . . . the mandate rule.”). In long-running litigation like this, the district court is especially constrained because it may not “do anything which is contrary to either the letter or spirit of the mandate” which we issued in our four previous remands. City of Cleveland v. Fed. Power Comm’n, 561 F.2d 344, 346 (D.C. Cir. 1977) (quoting Yablonski v. United Mine Workers of Am., 454 F.2d 1036, 1038 (D.C. Cir. 1971)). My colleagues do not discuss these two constraints, instead focusing on the “double dose of deference” they believe we owe the district court’s fourth-factor “reasonableness” assessment. Maj. Op. 6. Hence, they acknowledge our fourfactor test but do not apply it. See Maj. Op. 6 n.1 (“[W]e of course must and do adhere to our circuit precedent.”). In a telling footnote, they “doubt” that the first three factors have any role to play in “a proper interpretation of the statute.” Id. They suggest instead that “the fourth factor alone should constitute the test under FOIA for attorney’s fees.” Id. There may be good reason to question our FOIA precedent but, as a three-judge panel, we are bound to apply it.3 With respect, I 3 Some members of our court question the four-factor FOIA test and call for en banc review. See Morley VII, 719 F.3d at 690-91 9 believe the majority fails to do so. Indeed, the majority accepts that the first three factors favor Morley but does not review the district court’s reasoning and, worse, does not adequately evaluate the weight of the first three factors in light of Morley IX or Davy IV.4 See Maj. Op. 4-5. As a result, the majority necessarily relies on the “fourth factor alone” in affirming the district court’s determination that Morley is not entitled to attorney’s fees. Id.5 Under a faithful application of our fourfactor test, I believe the district court abused its discretion.6 (Kavanaugh, J., concurring); Davy IV, 550 F.3d at 1166 (Randolph, J., dissenting). 4 As discussed infra, it is not clear that the district court itself found that the first three factors favor Morley. 5 In attempting to establish that it does not rely only on the “reasonableness” factor of our test, the majority declares: “If the District Court had awarded attorney’s fees in this case, we would have affirmed.” Maj. Op. 14. This is pure dictum. The district court did not award fees and my colleagues’ declaration of what they would do in a hypothetical is entirely speculative. Moreover, had the district court awarded fees, there would have been no legal error to correct and no basis for remand. 6 Some of our opinions suggest that each of the four factors has a threshold that must be met. See Morley IX, 810 F. 3d at 844 (“[I]f it’s plausible ex ante that a request has a decent chance of yielding a public benefit, the public-benefit analysis ends there.”). Other opinions suggest that the inquiry is akin to a freestanding balancing test. See Cuneo, 553 F.2d at 1367 (“[T]he trial court must weigh the facts of each case against the criteria of the existing body of law on the award of attorney fees and then exercise its discretion in determining whether an award is appropriate . . . .”). In either event, the majority’s approach is flawed. If each factor can be met with by a “yes” or “no” answer, three in favor should outweigh one against. On the other hand, if the factors should be weighed against each other 10 A. Factor One: Public Benefit The district court found the potential public benefit of Morley’s request “small.” Its finding understates the importance of the Kennedy assassination.7 At least three times, we have recognized the potential public benefit of JFK-related FOIA inquiries. In Davy IV, we noted that the documents Davy sought provided “important new information bearing on the controversy over former District Attorney Jim Garrison’s contention that the CIA was involved in the assassination plot.” 550 F.3d at 1159 (quoting Davy Decl.) (alterations omitted). Then, in Morley VII, we vacated and remanded the district court order denying fees so that it could reconsider its publicbenefit analysis in light of Davy IV. Morley VII, 719 F.3d at 690. in a balancing test, the majority errs by failing to review fully the district court’s assessment of the first three factors. Few events in our national history have garnered as much attention as the assassination of President Kennedy. Three times since 1963, the Congress has investigated the details of the assassination. In 1964, the Warren Commission concluded that Lee Harvey Oswald killed President Kennedy alone and unaided. Joint Appendix (JA) 66. In 1978, however, the House Select Committee on Assassinations (HSCA) reopened the Kennedy investigation. JA 68. Ultimately, the HSCA concluded that Oswald had killed President Kennedy with unidentifiable co-conspirators; thereafter, the conspiracy theories multiplied. Id. In 1992, the Congress reentered the fray, enacting the President John F. Kennedy Assassination Records Collection Act of 1992, Pub. L. 102-526 (codified at 44 U.S.C. § 2107 Note) (JFK Act), which charged the Assassination Records Review Board to collect and release all unclassified documents related to the assassination. 7 11 Finally, we have expressly recognized the usefulness of Morley’s specific request. Morley IX, 810 F.3d at 844. Acknowledging that “a requester’s mere claim of a relationship to the assassination” does not “ipso facto satisf[y] the public interest criterion,” we noted that, if the subject of the request “is the Kennedy assassination—an event with few rivals in national trauma and in the array of passionately held conflicting explanations—showing potential public value is relatively easy.” Id. at 844 (second emphasis added). We continued: Morley’s request had potential public value. He has proffered—and the CIA has not disputed— that Joannides served as the CIA case officer for a Cuban group, the DRE, with whose officers Oswald was in contact prior to the assassination. Travel records showing a very close match between Joannides’s and Oswald’s times in New Orleans might, for example, have (marginally) supported one of the hypotheses swirling around the assassination. In addition, this court has previously determined that Morley’s request sought information “central” to an intelligence committee’s inquiry into the performance of the CIA and other federal agencies in investigating the assassination. Under these circumstances, there was at least a modest probability that Morley’s request would generate information relevant to the assassination or later investigations. Id. at 844-45. In other words, we held that Morley satisfied the public-benefit factor in this case. Id. at 844 (“[I]f it’s plausible ex ante that a request has a decent chance of yielding a public benefit, the public-benefit analysis ends there.”). 12 None of this is to say that Morley’s assassination theories necessarily have any merit. The point is that we have twice remanded the case based on the district court’s failure to assess properly the public benefit of Morley’s FOIA request. Thus, the district court’s description of the public value of the information sought by Morley as “small” ignores our decisions in Davy IV, Morley VII and Morley IX. See Kpodi, 888 F.3d at 491 (explaining mandate rule). B. Factors 2 and 3: The Requester’s Interest Factors two and three are controlled by Davy IV. See 550 F.3d at 1162 (“[T]he district court abused its discretion in determining that the second and third factors weighed against Davy . . . .”). In addressing these factors, the majority cites the oral argument transcript for the proposition that “the District Court ‘found that three of the four factors favored Morley.’” Maj. Op. 4 (quoting Oral Arg. Tr. at 4). Contrary to Morley’s counsel’s assertions, however, the district court did not hold that factors two and three favored Morley. Rather, the district court stated that “the first three factors do not clearly indicate whether the Court should award attorney’s fees—it is a very close call.” Morley X, 245 F. Supp. 3d at 78.8 Thus, it is far from clear how the district court viewed the second and third factors. If it believed the first three factors indeed favored Morley, the balance at that stage would have undoubtedly In a footnote, the district court provided a caveat: “In an abundance of caution, therefore, I will clarify that even if costs avoided do not count as a commercial benefit, the public interest in incentivizing Morley would be low enough in this case that I would still find the fourth factor dispositive.” Morley X, 245 F. Supp. 3d at 78 n.2. It is unclear how—or if—this comment affected the district court’s analysis but, in any event, the district court did not conclude that factors two or three affirmatively favored Morley. 8 13 tipped in his favor and there would have been no “tie” to break. Id. Moreover, this case is indistinguishable from Davy IV on factors two and three.9 Like Davy, Morley is a journalist. Like Davy, Morley “hope[s] to earn a living plying [his] trade” and receives modest remuneration for the articles he writes. Davy IV, 550 F.3d at 1160; see also Nationwide Bldg. Maint., Inc. v. Sampson, 559 F.2d 704, 712 (D.C. Cir. 1977) (“For the purposes of applying this criterion, news interests should not be considered commercial interests.”). Like Davy, Morley has an interest in investigating President Kennedy’s assassination. And like Davy, Morley may not be able to publish the information he obtains until long after his lawsuit ends. But unlike in Davy’s case—where we held that factors two and three favored Davy—the district court here held that the first three factors were a “tie” or a “very close call” despite already having counted the public benefit in Morley’s favor. Morley X, 245 F. Supp. 3d at 78. In my view, this holding was legal error. Nor does it make any difference that the CIA referred Morley to NARA rather than denying his request outright. Although the district court noted that Morley obtained some The majority suggests that I am “conflating our prior decision in Morley’s case and our prior decision in Davy.” Maj. Op. 13. Not so. I recognize that our mandate in Davy IV does not by its express terms apply to this case. Rather, Davy IV stands as legal precedent that defines the limits of the district court’s discretion to award fees. See Koon, 518 U.S. at 100. Hence, Davy IV is binding because it is factually on all fours with Morley’s case with respect to factors two and three. Indeed, neither the district court nor the majority attempts to distinguish Davy IV in analyzing the second and third factors. See Morley X, 245 F. Supp. 3d at 77-78 (citing Davy IV but failing to discuss its facts or its holding); Maj. Op. 5 (citing Davy IV with “cf.” signal and no accompanying explanation). 9 14 benefit by securing documents from the CIA rather than searching through NARA on his own, the record is clear that only 113 (of the 524) documents produced were available through NARA. Id. No amount of searching the public records would have unearthed those 411 documents. Unlike the plaintiff in Tax Analysts, who knew that the requested documents eventually would become public, 845 F.2d at 31819, Morley had no way of knowing whether the files he sought were available at NARA. See Davy IV, 550 F.3d at 1164 (Tatel, J., concurring) (“Before suing, requesters in Davy’s position have no idea what documents responsive to their FOIA requests might contain because the agency has told them nothing.”). In sum, factors two and three cannot be “close calls,” at least not after Davy IV. Davy IV makes clear that factors two and three unquestionably weigh in Morley’s favor and the district court erred in concluding otherwise. Davy IV, 550 F.3d at 1162. C. Factor Four: Reasonableness Finally, with regard to the fourth factor, “the heart of this case,” Maj. Op. 5, the majority—mistakenly, in my view— concludes that the CIA’s response to Morley’s request was reasonable. The fourth factor considers whether the agency’s opposition to disclosure “had a reasonable basis in law.” Tax Analysts, 965 F.2d at 1096. It examines whether the agency was “recalcitrant in its opposition to a valid claim or otherwise engaged in obdurate behavior.” Davy IV, 550 F.3d at 1162 (quoting LaSalle Extension Univ. v. FTC, 627 F.2d 481, 486 (D.C. Cir. 1980)). And the burden is on the CIA to show “that it had any colorable or reasonable basis for not disclosing the material until after [the plaintiff] filed suit.” Id. at 1163. In Morley II, we assessed the CIA’s response to Morley’s FOIA request and found it lacking. Specifically, we reversed 15 the district court’s summary judgment order and held that the agency’s response was legally insufficient on seven separate grounds. The majority discusses five in its opinion. It acknowledges that the CIA: (1) missed the 20-day statutory deadline to respond, Maj. Op. 7; (2) incorrectly referred Morley to NARA rather than responding to his FOIA request itself, Maj. Op. 7-11; (3) failed to search its operational files, Maj. Op. 11; (4) submitted an incomplete Glomar response, Maj. Op. 11-12; and (5) relied on an interpretation of Exemption 2 that was later overruled, Maj. Op. 12. It addresses these errors of law seriatim and labels them “incorrect legally,” Maj. Op. 9, but not “unreasonable.” To me, the CIA’s multiple flawed legal positions suggests that it was “recalcitrant” in declining to produce any documents before being sued. Davy IV, 550 F.3d at 1162. At the least, the errors collectively undermine the district court’s conclusion that the fourth factor “weighs heavily against Morley.” Morley X, 245 F. Supp. 3d at 78 (emphasis added). With respect to the CIA’s decision to refer Morley to NARA instead of producing any documents, however, I find the CIA’s—and the district court’s—positions entirely unreasonable. In Tax Analysts, we held that “in response to a FOIA request, an agency must itself make disclosable agency records available to the public and may not on grounds of administrative convenience avoid this statutory duty by pointing to another public source for the information.” 845 F.2d at 1067 (emphasis in original). We reaffirmed the holding in Morley II, declaring that “an agency has ‘withheld’ a document under its control when, in denying an otherwise valid request, it directs the requester to a place outside of the agency where the document may be publicly available.” Morley II, 508 F.3d at 1119 (quoting Tax Analysts, 492 U.S. at 150). 16 On its face, our holding in Tax Analysts (and the Supreme Court’s endorsement thereof) should control this case. It is undisputed that Morley made a valid FOIA request. It is likewise undisputed that the CIA initially directed Morley to “another public source for the information”—NARA—without producing any of the documents he requested. Tax Analysts, 845 F.2d at 1067. Thus, the CIA did not “itself” disclose its records to Morley. Id. Nonetheless, to bring this case within the ambit of Tax Analysts, the majority reasons that “Congress itself had provided ‘an alternative form of access’” to the records. Maj. Op. 9 (emphasis added). But the Congress is not the CIA and congressionally mandated access to documents is not the same as agency access under FOIA. Simply put, without statutory authorization, the CIA is not excused from its FOIA obligations. Both the district court and the majority use the JFK Act to support the reasonableness of the CIA’s initial withholding. The JFK Act instructs executive agencies to deliver documents related to JFK’s assassination to NARA for publication. As the majority notes, however, the statute also provides that “[n]othing in [the JFK] Act shall be construed to eliminate or limit any right to file requests with any executive agency or seek judicial review of the decisions pursuant” to FOIA. JFK Act, Pub. L. 102-526, § 11(b). The majority apparently reads this language to mean that the public may “file” a FOIA request but an agency has no duty to collect and produce documents it has already transferred to NARA. Maj. Op. 10. If the JFK Act ensures the public’s right to “file” a FOIA request, it necessarily preserves the agency’s duty to respond to that request. The right to file means little if the agency replies with nothing more than a letter. And we have so noted: “[section] 11(b) . . . provides that the [JFK] Act does not limit or eliminate any rights under FOIA.” Assassination Archives & Research Ctr. v. DOJ, 43 F.3d 1542, 1544 (D.C. Cir. 1995) (emphasis added). Other circuits agree with this common-sense interpretation of the JFK Act. See, e.g., Minier 17 v. CIA, 88 F.3d 796, 802-03 (9th Cir. 1996) (“[W]e hold that the JFK Act has no direct bearing on [a plaintiff’s] FOIA request.”). The CIA’s eventual document production here illustrates the difference between FOIA and the JFK Act. When Morley first made his request, neither he nor the CIA knew whether the documents he requested had been transferred to NARA. As it turns out, only 113 of the 524 documents were ever transferred. Morley X, 245 F. Supp. 3d at 76. If not for Morley’s lawsuit, the CIA never would have disclosed those non-transferred 411 documents. More to the point, neither statute justifies the CIA’s withholding the documents. Under these facts, I believe it was legal error to conclude that the CIA’s position was reasonable. See Koon, 518 U.S. at 100. In sum, I believe the district court erred on two levels: it erred in evaluating each of the four factors individually and abused its discretion in weighing them against one another. Accordingly, this case does not call for “[d]eference piled on deference.” Maj. Op. 11. It calls for an adherence to Davy IV and our four earlier Morley opinions. Because I believe the district court ignored our mandate and misapplied our precedent, I would vacate the district court order a fifth time and remand with instructions to award Morley the attorney’s fees to which he is entitled. Accordingly, I respectfully dissent.