This is the Justice Department’s Response to Order to Show Cause, filed July 2, 2026, in Phang v. Blanche, a federal case in the U.S. District Court for the District of Columbia before Judge Emmet G. Sullivan. To understand what that means, it helps to walk through the procedural chain that produced it. Assistance from Claude AI.
A little over a week before this filing, Judge Sullivan issued a 48-page memorandum opinion granting a preliminary injunction to plaintiff Katie Phang — a journalist who sued the Department of Justice for allegedly failing to comply with the Epstein Files Transparency Act (EFTA). A preliminary injunction is a court order issued before a case is fully litigated, meant to preserve the status quo or prevent ongoing harm while the underlying lawsuit continues. It is not a final judgment — a distinction that turns out to matter a great deal in this filing.
Rather than immediately ordering DOJ to unredact and release specific documents, Sullivan’s order took an unusual middle path: it directed the Department to either (1) take five specific actions — mostly releasing documents or unredacting names — or (2) show cause, meaning explain to the court’s satisfaction why it shouldn’t have to. This filing is DOJ’s attempt to do the latter: to justify its redactions and delays rather than comply outright.
Think of a show-cause order as the court saying, “I’ve seen enough to believe you may be breaking the law. Convince me otherwise, on the record, or I will make you fix it.” That is a meaningfully softer posture than a straightforward compliance order, and DOJ’s brief leans hard on that distinction throughout.
The Parties
Katie Phang is a journalist and legal commentator who sued the Department of Justice in April 2026, arguing that DOJ’s slow-walking of Epstein-related document releases was actively harming her ability to do her job — an “informational injury” theory that Judge Sullivan ultimately accepted. Her lawsuit does not proceed under the Freedom of Information Act (FOIA), the law most people think of when they picture records requests against the government. Instead, she sued under the Administrative Procedure Act (APA), arguing that DOJ’s noncompliance with EFTA amounted to unlawfully withheld agency action.
Todd Blanche, the defendant, is sued in his official capacity as Acting Attorney General — meaning the suit is really against the Department of Justice as an institution, not against Blanche personally. Blanche’s own position adds a layer of political texture to this case: he is a former personal defense attorney to President Trump who became Deputy Attorney General in 2025, was elevated to Acting Attorney General in April 2026 after Trump fired Pam Bondi, and was formally nominated in June 2026 to hold the job permanently. His Senate confirmation hearing is scheduled for July 15, 2026 — less than two weeks after this filing — and his handling of the Epstein files has already become a flashpoint in that confirmation fight.
The Law at the Center of the Case: What Is the EFTA?
Congress passed the Epstein Files Transparency Act in November 2025, and President Trump signed it into law the same day — a rare instance of a bill moving from passage to signature in a single day, reflecting the political pressure surrounding the issue. The law requires DOJ to make public “all unclassified records, documents, communications, and investigative materials” relating to Jeffrey Epstein, Ghislaine Maxwell, and the government’s investigations of them, across nine defined categories.
Crucially, the statute doesn’t grant DOJ unlimited discretion to withhold or redact material. It permits withholding only for five specific reasons: protecting victims’ personally identifiable information, protecting against disclosure of child sexual abuse material, protecting active investigations (narrowly and temporarily), protecting graphic images of death or abuse, and protecting properly classified national security information. Any redaction made under these grounds must come with a written justification published in the Federal Register — the government’s official daily journal for regulations, notices, and legal announcements — and reported to Congress.
The law also contains a pointed restriction: no record may be withheld “on the basis of embarrassment, reputational harm, or political sensitivity, including to any governmental official, public figure, or foreign dignitary.” Congress wrote that sentence, transparently, to prevent exactly the kind of selective redaction critics have accused DOJ of engaging in — protecting the reputations of well-connected people who appear in Epstein’s orbit.
How We Got Here
Phang’s core argument is not that DOJ released nothing. The Department made substantial releases beginning the day the EFTA took effect and continuing through January 2026, when it released the bulk of the files — technically weeks past the law’s 30-day deadline. Her argument is narrower and more specific: that among what DOJ redacted or withheld were names and documents that don’t fit any of the five permitted categories, and that DOJ never adequately explained why they were withheld, as the statute requires.
Judge Sullivan agreed, and agreed emphatically. His opinion repeatedly stated that Blanche had “conceded” DOJ was violating the Act — language DOJ now spends several pages trying to walk back, arguing the “concession” was really just a procedural default under a local court rule (more on that below, because it’s one of the more interesting threads in this filing).
The five categories of information Sullivan ordered DOJ to release or justify were: (1) eight emails with sender/recipient names redacted; (2) two documents — including a 2007 draft indictment — with potential “co-conspirator” names redacted; (3) underlying handwritten FBI interview notes behind four typed FD-302 interview reports; (4) foreign-language documents DOJ hadn’t yet reviewed for release; and (5) a formal redaction log for the Federal Register, as the statute requires.
The Government’s Core Legal Argument: You Shouldn’t Even Be Here
Before addressing any of the five specific categories, DOJ spends the first several pages of this brief re-litigating a threshold question Judge Sullivan already resolved against it: can a private citizen like Phang sue to enforce the EFTA at all?
This is where the brief gets into genuinely contested legal terrain, and it’s worth slowing down to explain the doctrine, because it’s the strongest card DOJ is holding.
The APA, private rights of action, and “adequate remedies”
Not every law that imposes duties on the government can be enforced by a private lawsuit. Generally, Congress has to create a “cause of action” — explicit legal permission for private parties to sue — for a statute to be privately enforceable. The EFTA doesn’t contain one. So Phang instead sued under the Administrative Procedure Act, a general-purpose statute that lets people challenge unlawful “agency action” in federal court.
But the APA itself contains a limiting principle, found in 5 U.S.C. § 704: judicial review under the APA is available only where there is “no other adequate remedy in a court.” DOJ argues FOIA — the well-established public records law — provides exactly that adequate remedy here, and that a body of D.C. Circuit and D.C. district court precedent holds that when a plaintiff is really just trying to get government records, FOIA forecloses an APA claim, even if the plaintiff might get a worse outcome under FOIA than under the more generous statute they’d prefer to invoke.
This is a real doctrine, not an invented one. DOJ cites CREW v. Department of Justice, Bowen v. Massachusetts, Electronic Privacy Information Center v. IRS, and several district court opinions, all standing for a consistent proposition: the “adequate remedy” bar doesn’t require an identical remedy, just one of the “same genre” — and document disclosure is document disclosure, whether sought under FOIA or otherwise.
Where this argument is genuinely strong: The case law DOJ cites is real, consistently applied within this circuit, and not fringe. Judges in the D.C. district have repeatedly turned away APA claims seeking documents on the theory that FOIA already occupies that field. If this were a garden-variety “I want records” case, DOJ’s argument would likely win.
Where this argument is genuinely weak: Judge Sullivan’s counter, buried in his opinion but paraphrased accurately in this brief, is the more interesting point: FOIA contains broad exemptions (for privacy, law enforcement records, and so on) that would likely let DOJ withhold much of what EFTA specifically compels it to disclose. If FOIA structurally cannot deliver the disclosure Congress mandated in EFTA, is it really an “adequate” substitute remedy — or just a remedy that happens to also involve documents? DOJ’s brief responds that “the possibility a plaintiff might fare worse under FOIA” doesn’t matter, quoting precedent to that effect — but that response somewhat begs the question when the “worse” outcome isn’t just less generosity, it’s a structurally different (and narrower) legal standard than the one Congress wrote specifically to override the usual secrecy defaults. This is a live, unsettled question, not one either side can claim is clearly settled by existing case law, because no prior case involved a disclosure statute as specifically drafted to override typical withholding grounds as the EFTA.
The “concession” fight
DOJ also spends real energy insisting it never conceded an EFTA violation — arguing that Sullivan’s characterization of a “concession” was actually just the mechanical effect of a local court rule (Local Rule 7(b)) that treats unaddressed arguments in a brief as conceded. Because DOJ’s opposition brief focused solely on jurisdictional arguments and never defended its redactions on the merits, the court treated the merits as unopposed — not because DOJ affirmatively admitted wrongdoing.
This is a fair technical point about civil procedure, and DOJ backs it with a genuinely applicable precedent: University of Texas v. Camenisch, a Supreme Court case holding that findings made at the preliminary injunction stage — where the parties haven’t had a full opportunity to litigate — shouldn’t bind a party for the rest of the case. That’s a correct statement of law and a legitimate distinction.
But it’s also a bit of a tightrope walk rhetorically. DOJ is arguing, in essence, “we didn’t concede violating the law, we just didn’t argue we hadn’t” — a distinction that is doctrinally real but likely to strike a skeptical reader (or judge) as a difference without much practical daylight. Sullivan already used blunt, repeated language in his opinion characterizing DOJ’s position as a concession. Convincing the same judge that he mischaracterized his own findings, one week later, without new facts, is a hard sell.
The Five Categories: Making the Factual Case, Document by Document
Having made its jurisdictional argument (largely as a hedge, since DOJ acknowledges “the Government understands the Court has ruled on this issue”), the brief pivots to a document-by-document defense of the specific redactions Sullivan flagged. This is where the analysis gets granular, and where the strength of DOJ’s position varies considerably by category.
1. The eight redacted emails
DOJ says six of the eight redactions are victim names — a category EFTA expressly permits withholding, and one the court itself called “appropriate.” That’s a reasonably strong, statute-grounded defense, assuming the factual claim holds up.
The seventh and eighth emails are different, and more interesting. DOJ says these redactions aren’t names at all, but private email addresses — which the Department treats as protected personal information under general Privacy Act principles, distinct from EFTA’s specific victim-protection carve-out. Here, DOJ makes an argument that cuts against itself: it notes that the identity of the person behind the seventh email — someone who told Epstein he “loved the torture video” — has already been publicly identified by members of Congress, including Rep. Thomas Massie, who used social media to name individuals after viewing unredacted files. DOJ argues that because the person is already publicly known, unredacting the email address “would serve no purpose beyond sharing that private email address.”
This is a genuinely awkward argument for DOJ to be making. If the person’s identity is already public — thanks to a sitting member of Congress broadcasting it — then the privacy rationale for continued redaction is considerably weakened, not strengthened. The Department is essentially arguing “everyone already knows, so there’s no point in us confirming it,” which is a practical point about redundancy, not a legal defense grounded in any of EFTA’s five withholding categories. If anything, it invites the response: if it’s already public, releasing it officially costs DOJ nothing and would demonstrate compliance goodwill. Relying on a congressman’s unofficial, out-of-process disclosure to justify continued official secrecy is a shaky foundation, legally speaking, and a court skeptical of DOJ’s good faith could easily read it that way.
2. Draft indictment and co-conspirator names
DOJ’s response here is more solid. Of four names redacted under a “co-conspirators” heading in a 2007 draft indictment, DOJ says three are actually victims (misfiled under a confusing label) and the fourth — Leslie Groff, an Epstein associate — has already been unredacted because she hasn’t claimed victim status. For a related document, DOJ says it cannot locate an unredacted version of that specific old draft, but has unredacted duplicates of substantially similar records confirming the redacted names are victims.
This is a plausible, fact-specific explanation, and if accurate, it’s a fairly persuasive one — it doesn’t ask the court to accept a policy argument, just a factual clarification that the “co-conspirators” label was misleading. Its main vulnerability is that DOJ is asking the court to take its word for factual claims (who is and isn’t a “victim,” which files are duplicates of which) without independent verification, at a moment when the court has already expressed skepticism about DOJ’s candor.
3. Underlying FBI interview notes
This is the weakest link in the brief, in this analyst’s assessment. DOJ declines to produce the handwritten notes underlying four FD-302 interview reports — including interview notes related to a woman who has accused President Trump of assaulting her as a minor, an allegation he denies — on the grounds that handwritten notes are “duplicative” of the final typed report and pose extra redaction risk because of their less standardized format.
The trouble is that “duplicative” is doing a lot of work here, and it’s an assumption, not a demonstrated fact. Interview notes and final reports are not identical documents by nature — investigators routinely leave detail, phrasing, and context in raw notes that gets summarized, condensed, or omitted in the polished report. That’s precisely why underlying notes are often independently significant in cases involving credibility disputes, which this is. EFTA’s text covers “records, documents, communications, and investigative materials,” language broad enough to plausibly include interview notes as distinct materials, not mere backups of the typed report. DOJ’s practical concern about redaction risk on handwritten material is a legitimate administrative burden, but it’s not one of EFTA’s five statutory grounds for withholding, and the brief doesn’t try to fit it into one — it simply asserts the notes are unnecessary. That’s an argument likely to draw real skepticism from a judge already primed to doubt DOJ’s redaction judgment.
4. Foreign-language documents
DOJ’s answer is essentially a resource-and-practicality argument: first-level reviewers couldn’t assess foreign-language documents for responsiveness, so DOJ told Congress that back in January, and “the Court should not supplant Congress’s acceptance of this explanation.” That last phrase is doing something legally unusual — DOJ is arguing that because it explained itself to Congress and Congress didn’t object, a federal court shouldn’t second-guess that explanation. But congressional silence isn’t the same as statutory compliance, and courts routinely evaluate agency compliance with statutes regardless of whether Congress has separately grumbled about it. This argument reads more like a plea for institutional deference than a legal defense, and it doesn’t squarely answer the court’s order, which asked DOJ to begin the review and report on its status — a comparatively modest ask that DOJ’s brief doesn’t clearly commit to.
5. Federal Register redaction log
Here DOJ essentially concedes the point and commits to compliance, noting it has already given Congress the required information and “will submit the notice for publication in the Federal Register as required.” This is the one item where DOJ isn’t fighting — a useful signal, since it suggests the Department is choosing its battles rather than resisting across the board.
The Evidence Marshaled
DOJ’s factual support comes almost entirely from its own prior communications: letters to Congress from December 2025 through February 2026, a “Section 3 Report” summarizing redactions, an internal “EFTA First Level Review Protocol” describing redaction procedures, and — notably — a sitting member of Congress’s social media posts and floor statements identifying individuals in the files. Using a lawmaker’s public commentary as evidentiary support for a redaction decision is an unusual move; it’s persuasive as a practical matter (it shows the information is genuinely already public) but it’s also an implicit admission that DOJ’s own redaction process has effectively been overtaken by informal disclosure happening outside the Department’s control — hardly a flattering picture of the process’s integrity.
The Precedents, and Whether They Hold
Beyond the “adequate remedy” cases discussed above, two other citations are worth flagging for what they reveal about DOJ’s strategy:
Youngstown Sheet & Tube Co. v. Sawyer (Justice Jackson’s concurrence, 1952) and Trump v. Mazars USA (2020) are both cited for the proposition that document disputes between branches of government are best resolved through political negotiation, not judicial intervention. Both are genuine, important separation-of-powers precedents — but both arose in a fundamentally different context: disputes between Congress and the Executive Branch over congressional subpoenas, not disputes between a private citizen and an agency over a disclosure statute Congress itself wrote to be automatically self-executing. Analogizing a citizen’s individual EFTA claim to an inter-branch subpoena fight is a stretch; it imports weighty constitutional language into a comparatively mundane statutory compliance dispute, and a skeptical judge may see through that framing.
Egbert v. Boule (2022) is cited for the general principle that only Congress creates causes of action — true as far as it goes, but Egbert is actually a case about refusing to extend implied constitutional damages claims against federal officers (a Bivens claim), a doctrinally distinct question from whether the APA’s general judicial review provisions cover EFTA noncompliance. It supports DOJ’s broader point but isn’t squarely on point.
Where the Government’s Case Is Strong, Where It’s Weak
Strongest ground: The jurisdictional/adequate-remedy argument is a real, doctrinally grounded position with supportive circuit precedent, even if it faces a serious counterargument specific to EFTA’s unusual structure. The Federal Register compliance item is essentially uncontested. The victim-PII redactions in the majority of the eight emails, and the co-conspirator clarifications in the draft indictment, are plausible, fact-based defenses squarely within EFTA’s permitted withholding categories.
Weakest ground: The FBI interview-notes argument doesn’t fit any statutory withholding category and rests on an unverified assumption of duplicativeness. The foreign-language-documents response doesn’t really answer what the court asked. And the reliance on a congressman’s public identification of an email’s author to justify not officially confirming that identification is a rationale that, read closely, undercuts itself — it’s a practical observation dressed up as a privacy defense.
There’s also a broader rhetorical vulnerability running through the brief: DOJ spends its opening pages insisting it has “not knowingly violated, nor has it ever acknowledged violating” EFTA, while several pages later making individualized, document-by-document arguments that only make sense if specific redactions actually were made outside the statute’s permitted grounds (like the co-conspirator mislabeling, since corrected). A brief that simultaneously denies any violation while walking back several specific redactions reads, to a skeptical eye, less like vindication and more like selective damage control.
Who Prevails?
This is not a case where the court is choosing between two evenly balanced positions on a blank slate — Judge Sullivan has already ruled once, in detailed and unusually pointed language, against the Department’s overall position. Show-cause proceedings exist precisely because the court has provisionally decided against a party and is giving that party a last chance to change its mind. Historically, that’s a difficult posture to reverse, especially before the same judge, absent significant new facts or law — and DOJ doesn’t present much of either here.
My assessment: DOJ is likely to succeed, at least partially, on the narrow, fact-specific redactions it can tie directly to EFTA’s five permitted categories — the victim-name redactions and the corrected co-conspirator labeling most plausibly survive scrutiny, because they map cleanly onto statutory text the court has already said it respects. DOJ is considerably less likely to succeed on the FBI interview notes and the foreign-language document review, where its arguments rest on administrative convenience and deference-to-Congress theories rather than the statute’s actual withholding grounds. The jurisdictional argument, while doctrinally serious, is unlikely to persuade the same judge who already rejected it once on a fuller record than what’s presented here — though it may well be the argument DOJ ultimately takes to the D.C. Circuit on appeal, where a fresh panel would evaluate it without the baggage of Sullivan’s prior findings.
Overall, expect Judge Sullivan to narrow but not eliminate his order — likely accepting DOJ’s explanations on the majority-victim-PII emails and the co-conspirator identifications, while pressing DOJ further on the interview notes and foreign-language review, and reiterating that the jurisdictional argument is settled for purposes of this case.
What Happens Next
For DOJ, the near-term choices are: comply with whatever the court orders on the remaining contested items, or seek a stay and pursue an emergency appeal to the D.C. Circuit. The brief explicitly hedges on this, asking for a 60-day stay to let the Solicitor General decide on appeal, or at minimum a 7-day stay to make that determination — language that signals real internal uncertainty about whether an appeal is worth pursuing, and on what timeline. Complicating that calculus: Blanche’s Senate confirmation hearing is set for July 15, meaning any continued public fight over Epstein-related redactions will play out just as senators are preparing to question him directly on this exact issue.
For Phang, the next step is almost certainly opposing any stay request and continuing to press for full compliance, likely through further briefing or a request that the court promptly rule on the show-cause response. Because this remains a preliminary-injunction posture rather than a final judgment, the underlying merits litigation continues regardless of how the show-cause dispute resolves — meaning this fight over specific documents is unlikely to be the last word in the case.
Broader Implications
This case sits at an unusually dense intersection of transparency law, separation-of-powers doctrine, and live political controversy. A few threads worth watching:
For administrative law generally, if Judge Sullivan’s adequate-remedy reasoning survives appeal, it could open a path for private enforcement of other disclosure statutes that, like EFTA, don’t contain explicit private rights of action but do contain specific, non-discretionary release mandates that FOIA’s exemptions wouldn’t otherwise honor. That would be a meaningful expansion of who can sue whom over government transparency failures — precisely why DOJ is fighting the jurisdictional point so hard, independent of the merits of any individual redaction.
For the Epstein files saga specifically, this filing underscores how contested the redaction process has become, with a sitting member of Congress effectively conducting his own informal declassification process on social media faster than DOJ’s official one — a genuinely unusual dynamic that raises real questions about the coherence and control of the Department’s compliance effort.
For the confirmation fight, this filing becomes part of the public record just as Blanche heads into hearings where his handling of exactly these documents is expected to be a central line of questioning, including from senators in his own party who have already expressed skepticism about aspects of his tenure. A federal judge’s finding that Blanche’s Department “conceded” violating a transparency law Trump signed under political pressure is not the kind of finding a nominee wants circulating two weeks before his hearing, regardless of how DOJ characterizes the technical basis for that finding in a subsequent brief.