House Oversight’s November 2026 Epstein Report: Findings, Evidence and Unanswered Questions

On November 10, 2026, the Republican majority staff of the House Committee on Oversight and Government Reform released a 137-page report on the federal government’s handling of the Jeffrey Epstein and Ghislaine Maxwell investigations. It brings together congressional interviews, earlier Justice Department investigations, financial records, correspondence and survivor accounts. Its central argument is that weak investigations, an unusually favorable prosecution agreement, institutional deference to wealth, and failures to protect survivors allowed abuse to continue. Official release.

The report is also a political document. It criticizes federal prosecutors, banks, prominent associates and the Justice Department’s disclosure process while defending President Donald Trump and attacking congressional Democrats. Some conclusions are supported by earlier official findings; others are allegations, disputed interpretations or claims that go beyond their cited sources. Comparing the report with the underlying records reveals substantive errors, including its claim that Lesley Groff never completed a law-enforcement interview. Majority report, pp. 1–7; transcript, p. 180.

This article follows every major chapter, explains the evidence behind its principal claims, and distinguishes congressional conclusions from testimony, regulatory findings and court decisions. References to the majority report use the printed page number. The PDF viewer’s page number is one higher because of the cover. A witness’s appearance in a directory, photograph, flight record or financial transaction is not by itself evidence of participation in abuse.

Read the report

Open or download the complete report · Official House copy · House release and transcript collection

Contents

• 1. What the report covers and how to read it
• 2. The investigation and chronology
• 3. Recruitment, coercion and the roles of assistants
• 4. The Florida investigation, non-prosecution agreement and work release
• 5. Wealthy clients, staff, banks and institutional access
• 6. International travel, recruitment and intelligence claims
• 7. Custody failures and Epstein’s death
• 8. Disclosure, survivor privacy and the political dispute
• 9. Findings and all six recommendations
• 10. Material source corrections and unresolved questions
• 11. Interview and deposition source index
• 12. Related Wiki coverage and community sources

1. What the report covers and how to read it

The document’s full title is A Review of the Federal Government’s Mishandling of the Jeffrey Epstein and Ghislaine Maxwell Criminal Investigations. It was issued under Chairman James Comer as a majority staff report, not a bipartisan committee judgment or a judicial ruling. The committee describes a 15-month investigation involving 19 completed depositions and transcribed interviews, document subpoenas, survivor meetings and financial-record review. Its published schedule separately lists Leon Black’s September 3, 2026 nonappearance. Majority report, pp. 8–10; Majority report, pp. 115–116.

The report’s strongest recurring theme is how ordinary-looking arrangements could sustain control: a job, an apartment, an educational opportunity, a travel booking, a bank account or an introduction. The underlying question is who knew what those arrangements were being used for, what warning signs were available, and what each institution did in response. Evidence of an arrangement and evidence of knowing criminal participation are different matters.

• Established historical outcomes: Epstein pleaded guilty to Florida state offenses in 2008. He died before trial on the 2019 federal charges. Maxwell was convicted in 2021 and sentenced in 2022. The 2019 indictment’s allegations must not be described as a verdict against Epstein. Majority report, pp. 13–14.
• Congressional credibility judgments: the majority rejects several witnesses’ accounts of ignorance. Those are staff assessments, not perjury convictions or adjudicated findings against every person discussed.
• Regulatory and investigative findings: OPR, OIG and NYDFS examined different questions under different standards. Their actual conclusions sometimes differ materially from the majority’s shorthand.
• Survivor accounts: testimony can explain coercion and institutional failure without publishing private identifiers, graphic images or identifying details removed from source records.
• Negative evidence: a witness’s lack of knowledge limits what that witness can establish. It neither proves a suspected crime nor categorically disproves conduct outside the witness’s knowledge.

2. The investigation and chronology

The historical account begins with Palm Beach police in 2005, the federal investigation in 2006, the September 2007 non-prosecution agreement, and the June 2008 state guilty plea. Epstein was released from jail in July 2009. Federal authorities arrested him in New Jersey on July 6, 2019; he was detained at MCC in New York and died on August 10. Maxwell was arrested in July 2020, convicted in December 2021 and sentenced to 20 years in June 2022. Majority report, pp. 13–14; OPR executive summary.

The congressional chronology includes February and May 2025 requests for DOJ records; August 2025 subpoenas to DOJ, former officials and the Epstein estate; September and November estate releases; November subpoenas to JPMorgan and Deutsche Bank; and a request to the U.S. Virgin Islands. The report says its process produced more than 65,000 publicly released pages, while the staff reviewed hundreds of thousands of documents. Received records, reviewed records and publicly released records are different inventories. Majority report, p. 10; Majority report, pp. 14–18; Majority report, pp. 115–116.

The Epstein Files Transparency Act became law on November 19, 2025. Its 30-day deadline for covered records fell on December 19. DOJ’s later January 30, 2026 release was measured in pages, not millions of individual files: more than three million additional pages, bringing the cumulative total to nearly 3.5 million pages. The majority’s timeline and later discussion blur those units. Public Law 119-38; DOJ’s January 30 announcement; Majority report, p. 17; Majority report, p. 122.

Black appeared voluntarily in June 2026 but declined to answer some questions. The committee then subpoenaed further testimony and documents. His later nonappearance led to committee and House contempt actions in September. The House resolution concerns compliance with congressional subpoenas; it is not a conviction for sexual misconduct. The report also records a planned Alan Dershowitz interview that did not occur. Majority report, p. 18; H.Res.1543, agreed September 16, 2026.

3. Recruitment, coercion and the roles of assistants

Massages, schedules and the directory

Chapter I describes the use of purported massage appointments to bring women and girls into Epstein’s residences. Groff’s testimony supplies operational detail: Epstein directed whom she should call, staff kept a computerized directory, and printed copies became the book commonly called the “Black Book.” She believed Maxwell created the original directory and acknowledged adding information. The majority reproduces massage-contact pages and points to descriptors such as “be discreet.” The directory helps document an administrative system; it is not a list in which every name is a perpetrator or victim. Majority report, pp. 19–24; transcript, pp. 30–36.

Groff described massage scheduling as routine, said she did not check ages or professional qualifications, and denied knowing that appointments involved abuse. She said she worked in separate offices before about 2013 and had not seen the massage room. The majority rejects her account and recommends renewed questioning. An accurate summary must preserve both the admitted scheduling activity and her denial of knowledge, rather than substitute the majority’s disbelief for an adjudicated finding. transcript, pp. 43–44, 211–212; Majority report, pp. 20–24.

Sarah Kellen’s account of dependency and abuse

Kellen is central to the report’s account of coercion. In her May 21, 2026 interview she described being recruited while financially insecure in Hawaii, offered work with Epstein and Maxwell, introduced to extraordinary wealth, and then required to remain constantly available. She said she initially worked without pay and that pay began only after Epstein demanded sexual submission. She described repeated abuse, intimidation, restricted autonomy and fear that his connections made escape impossible. These are her accounts of her own experience, not statements that every assistant had the same experience. Majority report, pp. 24–30; Kellen transcript, pp.9–13, 18–29.

The report follows several mechanisms of control separately. Kellen said Epstein chose where she lived and repeatedly promised, without delivering, to transfer an apartment into her name. She described clothes and hair being controlled, debt relief and gifts creating obligations, and a design business whose only client was Epstein. Lawyers and accountants handled the business, while Epstein determined staffing. What looked like professional independence could therefore leave the same dependency intact. Majority report, pp. 37–41; Kellen transcript, pp.23–24, 75–79, 119.

Kellen also said encounters with political leaders, royalty and wealthy clients reinforced her belief that resistance would be futile. That testimony concerns the effect of visible access on her. It does not establish that every person she saw knew of the abuse, facilitated it or participated in it. Similarly, housing, tuition and gifts are relevant to a coercion account when their conditions are demonstrated; their existence alone does not establish that every recipient was trafficked. Majority report, pp. 38–41.

Her account of victimization exists alongside a contested legal history about her own role. The 2007 agreement named her as a potential co-conspirator. In Maxwell’s case, evidence about Kellen’s scheduling and subordinate role supported the leadership enhancement upheld by the Second Circuit. Those findings arose in Maxwell’s case, not a criminal conviction of Kellen. The majority’s simple division between abused “personal assistants” and administrative enablers does not resolve every question about overlapping vulnerability, conduct and responsibility. United States v. Maxwell, September 17, 2024, pp.24–25; Majority report, pp. 42–44.

Maxwell and Groff had different evidentiary and legal positions

The majority describes Maxwell as both an organizer and a direct participant in abuse, drawing on survivor testimony and the criminal case. It describes Groff’s scheduling, communications, travel arrangements and housing coordination as logistical support. The two cannot be treated as legally interchangeable: Maxwell was convicted; the majority’s allegations about Groff’s knowledge are congressional conclusions that she disputes. Groff also distinguished her work from traveling with Epstein, although she described separately visiting his properties. Majority report, pp. 42–44; transcript, p. 61.

The Victoria’s Secret representation

Kellen said Fekkai introduced Epstein to her as a Victoria’s Secret scout, leading to a supposed casting in Los Angeles. She expressly distinguished that initial encounter from later assaults, saying he did not sexually assault her during that first casting. Her claim is that the misleading professional representation created access and credibility. The report connects this to Epstein’s financial relationship with Wexner. Majority report, pp. 30–33; Kellen transcript, pp.10–11, 17–18.

Wexner acknowledged hearing that Epstein was presenting himself as a scout, calling him and accepting his denial. His testimony at first described one occasion; later questioning elicited an account that may have involved a separate complaint. His uncertainty should remain visible rather than be converted into a precise chronology. Wexner denied authorizing Epstein to recruit models and said he would have dismissed him had he believed the misconduct. The majority criticizes his reliance on Epstein’s denial as inadequate scrutiny. transcript, pp. 94–96, 143–145; Majority report, pp. 31–33.

Allegations against Philip Levine and Frédéric Fekkai

Kellen alleged that Fekkai assaulted her in Maui after offering an opportunity connected to a hair show, and that Levine assaulted her while she was staying with Epstein and Maxwell in Saint-Tropez. Levine was mayor of Miami Beach, not Miami, from 2013 to 2017. A June 4, 2026 committee letter asked DOJ to investigate the allegations. A referral requests investigation; it does not establish guilt or that prosecutors have filed charges. Majority report, pp. 26–27; Majority report, pp. 31–32; June 4 referral.

Both men denied the allegations. Fekkai’s representative denied assault and denied introducing her to Epstein. Levine’s spokesperson described a consensual encounter and disputed any contrary allegation. Their denials belong beside the allegations, not in a distant source list. CNN’s June 4 report, including the responses.

Modeling agencies, introductions and princess references

The report discusses Jean-Luc Brunel, MC2 Model Management, Daniel Siad, and Barbro Ehnbom’s network as different routes through which Epstein met young women. It cites modeling photographs, housing, financial support and meeting arrangements. Groff acknowledged sending photographs and reserving apartments when directed, while denying knowledge of abuse. The majority infers coordinated exploitation from these patterns; the existence and extent of a criminal agreement must be established individually, not assumed from introductions. Majority report, pp. 33–34; transcript, pp. 69, 105, 111, 118.

The executive summary incorrectly calls Brunel a convicted sex trafficker. Brunel was charged in France and died in custody in 2022 before trial; he denied wrongdoing. Ehnbom told SVT in April 2026 that the meetings she organized concerned education and careers and that she knew of no introduced woman who had been abused. Those facts do not eliminate legitimate questions about continued contact after Epstein’s conviction, but they prevent the majority’s interpretation from becoming an unsupported finding against everyone involved. Majority report, pp. 4–5; Brunel’s death before trial; Ehnbom’s SVT interview.

Pages 34–36 interpret “Snow White,” “Beauty and the Beast” and other fictional-character references as names used for women. The majority juxtaposes a woman’s message, Staley’s “Say hi” email and other records, and rejects his explanation. Staley denied sexual contact with a person dressed as Snow White and said he might have encountered her at Epstein’s apartment. The existence of coded or suggestive language can warrant investigation, but it does not by itself identify every participant, establish an individual’s age, or prove an assault. There is no need to reproduce intimate photographs or identify the redacted woman to explain the evidentiary dispute. Majority report, pp. 34–36; transcript, pp. 31–36, 52–53.

4. The Florida investigation, non-prosecution agreement and work release

The bargain and its limits

Federal prosecutors prepared a proposed 60-count indictment and prosecution memorandum in 2007, but no such indictment was filed. Instead, the Southern District of Florida signed a non-prosecution agreement on September 24, 2007. It required a state guilty plea, incarceration, sex-offender registration and arrangements intended to facilitate civil recovery. The bargain also protected four named potential co-conspirators and other potential co-conspirators within its applicable scope. It was a prosecutorial commitment, not a conviction of those named or unlimited nationwide immunity for future conduct. The Second Circuit later held that it did not bind New York prosecutors in Maxwell’s case. Majority report, pp. 45–46; OPR executive summary, pp.i–iii; Maxwell appellate decision.

The majority treats the bargain as the decisive lost opportunity to stop subsequent abuse. OPR’s 2020 investigation had already criticized resolving the case before important investigative work was complete, insufficient consultation, and failures in the treatment of victims. OPR found poor judgment but did not find professional misconduct by the prosecutors it examined or evidence that Acosta’s resolution was based on corruption or impermissible considerations such as wealth and status. The majority’s stronger later claims must remain distinct from those disciplinary findings. Majority report, pp. 45–46; Majority report, pp. 53–63; DOJ’s November 2020 account of OPR’s conclusions.

Secrecy affected different people at different times. OPR distinguishes disclosure to litigating victims in 2008 from public release after a state court order in September 2009, nearly two years after execution. Kellen’s account of learning much later that she was named is a separate personal chronology. The report’s shorthand about roughly a year of secrecy does not capture those distinctions. Majority report, pp. 45–46; OPR executive summary, pp.ii–iii.

Acosta’s oversight and the conflict allegations

The majority criticizes Acosta for delegating responsibility and for gaps in his later recollection. He accepted responsibility for the agreement he approved. An inability to remember a particular event in a 2025 interview does not establish what he did or did not know in 2007. OPR’s examination of contemporaneous records documented substantive emails, review of drafts and approval, while still criticizing the process and judgment. Majority report, pp. 46–48; OPR report, pp.174, 182–183, transcription of court-filed report.

The report describes Matthew Menchel’s earlier relationship with defense lawyer Lilly Ann Sanchez as a potential conflict and suggests it may explain favorable treatment. OPR found that disclosure would have been prudent so supervisors could assess appearances. It also expressly found no evidence that Menchel instigated the plea as a favor to Sanchez. That limiting finding appears on the same OPR page used by the majority and must accompany the suspicion. The report’s later-meeting and call records do not, on their own, establish a corrupt bargain during his government service. Majority report, pp. 48–50; OPR report, pp.153–155 and note 226.

The skiing-photograph allegation is particularly uncertain. The majority says it has not independently verified either the photograph’s existence or authenticity, and Menchel denied skiing with Epstein. Acosta’s reaction to a question about the alleged image is not corroboration. The report consequently cannot support stating that a verified photograph proves an undisclosed friendship. Majority report, pp. 48–49; Miami Herald, including Menchel’s denial.

For Bruce Reinhart, the majority discusses his move from the U.S. Attorney’s Office into private practice, representation of Epstein associates and business-address records. Kellen described Epstein choosing and paying a lawyer for her civil deposition and said she regarded the lawyer as aligned with his interests. This is testimony about her experience, not a judgment that Reinhart negotiated the NPA or illegally influenced the charging decision. Majority report, pp. 50–53.

A 2013 government response to requests for admission supplies more precise context: the government admitted that Reinhart learned confidential, nonpublic information and discussed the matter with another AUSA, but denied that he participated in decision-making. Those admissions do not establish that he transmitted information to Epstein or participated in a corrupt agreement. Shared offices and incorporation dates also do not prove that every person associated with an address participated in misconduct. Government admissions, request 15, Document 213-1 p.9, refiled as 361-31.

Missing computers, investigative opportunities and the Berger theory

OPR documented significant unfinished work: missing computer equipment, unresolved efforts to obtain it, additional victims who had not been interviewed, and incomplete attempts to secure cooperation. Investigators believed the computers might contain material that could corroborate accounts or strengthen a federal case. Missing equipment is an established investigative gap; its supposed contents are not a completed forensic finding. Majority report, pp. 53–58; OPR report, pp.45–47, 175–177.

Kellen later recalled overhearing instructions to retrieve computers but could not date the call, establish that the instruction was carried out or identify where equipment went. Her account of collecting directories and printed photographs for lawyers concerned the Manhattan residence around 2006 and should not be merged with the October 2005 Palm Beach search. Indyke’s account of learning about hard drives in private investigators’ possession was also retrospective. Majority report, pp. 55–57.

The report raises a theory about Sandy Berger by combining an urgent phone message before the Palm Beach search with a later deposition question. Timing and a question’s premise do not establish the content of a conversation or a tip-off. The report itself leaves the question unresolved. A refusal to answer under a claimed privilege is not independent proof that the premise was true. Majority report, p. 58.

Interview failures: what the records actually show

The majority argues that too many well-placed witnesses escaped substantive questioning. Its concern about incomplete follow-up deserves examination, but its description of Groff is demonstrably wrong. Page 61 says she was never successfully interviewed by law enforcement and cites transcript page 180, lines 10–16. In those exact lines, Groff confirms a completed voluntary FBI interview in 2021 and says she was not interviewed on another occasion. The questioning calls it a September 24 interview, but the same Bates-numbered FD-302, “Proffer of Lesley Groff,” EFTA01246216–EFTA01246224, records the interview as July 23, 2021, the drafting date as August 3 and the entry date as September 24; page 2 repeats July 23. This is a date discrepancy in the descriptions of the cited record, not evidence establishing a second interview. Her lawyer clarifies that the FBI statement was not sworn. Majority report, pp. 59–63; transcript, p. 180.

The chronology also includes an attempted FBI interview at her home in August 2007, which she declined without counsel. She disputed the government document’s account of whom she telephoned, saying it was Indyke rather than Epstein. Kellen described an attempted FBI approach at the St. Thomas airport in 2007. “No contact,” “an attempted interview” and “a completed substantive interview” are different claims. OPR also records interviews with numerous employees, so a later-selected witness list cannot establish that agents interviewed nobody in the organization. transcript, pp. 90, 180, 218; Majority report, pp. 60–63; OPR report, p.175 note 252.

The majority also criticizes a geographically narrow investigation in light of allegations and residences outside Florida. That is an institutional-coordination issue. A congressperson’s question containing an allegation or chronology is not itself sworn corroboration of those premises, particularly where the witness declines to adopt them. Majority report, pp. 58–63.

Work release and continued vulnerability

The report describes work release of up to twelve hours a day, six days a week; the Florida Science Foundation office; privately financed off-duty deputies; weak monitoring; and visits to Epstein’s home. Acosta said prosecutors had received assurances of confinement and regretted relying on the state system. The agreement, however, did not expressly prohibit work release. A harmful discretionary decision, failure to enforce a bargain and a proved criminal offense are different conclusions. Majority report, pp. 63–66; OPR executive summary, p.iii.

Kellen’s testimony and civil complaints cited by the report allege continued abuse during the custody period and work release. Those accounts explain why the conditions of confinement matter beyond sentencing optics. They should be identified as testimony and allegations, rather than trial findings about every incident, and can be explained without graphic detail or identifying anonymous complainants. Majority report, pp. 64–66.

OPR records a November 24, 2008 notice to defense counsel that prosecutors considered work release a breach, followed by no further enforcement action. The agreement’s wording and concessions complicated enforcement. This supports criticism of an ineffective bargain. It does not establish the majority’s much broader proposition about what every DOJ or FBI employee did or considered throughout the following decade. The new 2019 New York prosecution was a separate development. Majority report, p. 66; OPR report, pp.115, 182.

Victims’ rights and the omitted appellate ruling

The report cites the 2019 district-court decision finding violations of the Crime Victims’ Rights Act in the handling of the agreement, but omits the later controlling procedural outcome. In April 2021, the en banc Eleventh Circuit denied Courtney Wild’s mandamus petition, holding that the statute did not authorize this standalone civil enforcement action without an existing criminal proceeding. The court condemned the treatment of victims while rejecting the asserted judicial route. The Supreme Court denied review in February 2022. Majority report, p. 67; In re Wild, en banc opinion, pp.2–3, 12–13; Supreme Court docket 21-351.

The later ruling neither endorsed the treatment of survivors nor invalidated the NPA. The legal history therefore requires both the district-court finding and the subsequent appellate limit. Acosta’s explanation that notification might expose victims to attacks about financial motives does not erase OPR’s criticism of the lack of forthright, sensitive treatment. OPR executive summary.

NDAs and Black’s testimony

The majority treats confidentiality agreements as a further obstacle to disclosure. It cites Black’s testimony about a private relationship, an NDA and Epstein’s advice. The underlying transcript is narrower than the report’s assertion that Epstein advised on the NDA: Black denied that Epstein negotiated it and said he discussed a claimed extortion problem and estate implications. Counsel later denied Epstein’s involvement in any NDA. The claim of extortion is Black’s allegation, not a finding against the private woman. Majority report, pp. 67–68; transcript, pp. 41–43, 47.

The NPA’s secrecy, lack of victim notification and private NDAs are separate legal mechanisms. None should be used as a shortcut to conclude that every confidential settlement concealed trafficking. The policy issue is whether agreements deter people from reporting sexual abuse, and whether survivors have a practical route to speak and seek remedies.

5. Wealthy clients, staff, banks and institutional access

Wexner’s authority and the limits of the financial account

Chapter III argues that prominent relationships made Epstein appear legitimate to new clients and institutions. Wexner’s testimony documents unusually broad authority, including a 1991 power of attorney and a foundation role. Wexner said that discovering alleged theft and learning of criminal conduct led him to end the relationship. He denied contact after 2008. His account of financial losses remained incomplete: he said he later understood the loss to be at least $100 million but did not know the full total or how it was accomplished. That is an attributed claim, not a completed forensic accounting. Majority report, pp. 69–73; Majority report, p. 77; transcript, pp. 76–85, 106.

The report sometimes treats Epstein’s own claims about famous clients as established relationships. Wexner’s testimony and counsel’s interventions make clear that some information came from Epstein’s representations. A named client, claimed client, proposed engagement and executed agreement should not be combined into one verified roster. transcript, pp. 76–77.

Kahn’s client list and Black’s payments

Kahn identified five sources of fees: Wexner; Glenn Dubin or Highbridge; Steven Sinofsky; Ariane de Rothschild or an affiliated entity; and Leon Black. He qualified his knowledge. He recalled roughly $15–25 million connected with Highbridge’s sale, believed Sinofsky paid $5 million without knowing the work performed, and recalled roughly $20–25 million associated with Rothschild estate planning. The list is his account, not a finding that each person knowingly financed crimes or that every amount is independently reconciled. Majority report, pp. 72–76; transcript, pp. 43–47.

Black described substantial tax and estate-planning payments, introductions to prominent people and reliance on the Apollo-commissioned Dechert review. He denied knowing of or participating in Epstein’s criminal conduct. He cited $158 million in fees and a separate $30 million loan. In March 2025, Senate Finance investigators reported records supporting $170 million in payments, $12 million above the Dechert accounting. These are differing attributed accounts, not interchangeable measures. Loan principal, service fees, claimed tax savings and settlements cannot be added indiscriminately. Majority report, pp. 73–78; transcript, pp. 14–15, 29–31; Senate Finance, March 12, 2025.

The majority questions the review’s independence by raising lawyer Andrew Levander’s earlier contacts with Epstein. That is a concern advanced by the majority, not an adjudicated conflict finding. Black said he did not know the nature of the earlier relationship. Separately, Black agreed to pay $62.5 million in a January 2023 settlement with the Virgin Islands. The agreement’s recitals connect his payments to Epstein’s operations, but the settlement is not a conviction or an admission that Black intentionally funded abuse. Its own limits on use as evidence should not disappear from the account. transcript, pp. 29–34; Executed USVI settlement, p.1.

Ruemmler, Lutnick, Gates and Staley

Kathryn Ruemmler described Epstein introducing her to Edmond de Rothschild Bank, which became her legal client. She acknowledged gifts and a first-class travel upgrade but said she rejected a $50,000 private-jet card. Her explanation of a May 2019 legal discussion was that it concerned a threatened civil claim, not knowledge of the contemporaneous criminal investigation. Gifts and familiarity can be relevant to judgment and dependency without establishing obstruction. The report’s suggestion of compromised professional judgment should be identified as its inference. Majority report, pp. 78–81; transcript, pp. 25, 77–89, 122–123, 143–144.

The Ruemmler passage also contains citation and dating problems. One footnote reuses a file identifier described elsewhere as a visa email for a dinner-related claim. The transcript describes a May 6 email arranging a meeting the next day, so the email’s date should not automatically become the meeting date. Footnotes that print an Avenatti tweet in 2029 conflict with the transcript’s 2019 chronology. These inconsistencies call for checking the cited item, not reconstructing a narrative from the citation alone. Majority report, p. 75; Majority report, pp. 79–80; transcript, pp. 122–123, 143.

Howard Lutnick described limited encounters over fourteen years as a neighbor, including a 2011 foyer encounter and a late-2012 family lunch on Little St. James. The committee examines those contacts and their relationship to his public account. The presence of a family at a lunch establishes a visit, not participation in sexual misconduct. Majority report, pp. 82–83; transcript, pp. 9–10, 22–23.

Bill Gates said meetings beginning in 2011 concerned philanthropy, denied being a financial client, and said direct contact ended in late 2014. He also described later incoming communications and an indirect refusal to reimburse expenses related to a personal relationship. That is more nuanced than a claim that no communication whatsoever followed 2014. He denied that he was blackmailed while describing messages he interpreted as suggesting Epstein considered pressure. An unexecuted document does not establish an actual client engagement. Majority report, pp. 83–85; transcript, pp. 20, 93, 127–128.

Staley acknowledged that Epstein’s introductions had commercial value to JPMorgan and that the bank supplied credibility to Epstein. The “more than $100 million” figure he discussed was revenue from referred clients, learned from later reading. It was not direct fees paid by Epstein or a personal referral payment to Staley, which he denied. The report’s broader theory of mutually reinforcing prestige is an interpretation of those relationships, not proof that all revenue was criminal. Majority report, pp. 85–86; Majority report, p. 93; transcript, pp. 86–88.

Paid staff, loans and everyday administration

The report reconstructs Epstein’s operations through legal services, bookkeeping, cash delivery, scheduling and housing. Indyke described approximately $30 million of compensation over 23 years; Kahn described substantial loans and handling payments and cash; Groff described benefits and administrative work. Compensation and dependence are relevant evidence of incentives, but they do not establish that every employee knew every payment’s ultimate purpose. The majority’s judgments about loyalty and deliberate blindness need that distinction. Majority report, pp. 86–92; transcript, pp. 73, 91; transcript, pp. 75–86.

The report says Indyke offered no justification for repeated sub-threshold withdrawals. His transcript does offer one: he said $7,500 was a bank-imposed third-party withdrawal limit, that cash met household and petty-cash needs, and that he did not intend to evade reporting. He later claimed he alerted bank staff when the combined amount could exceed $10,000 and invited appropriate reporting. His account does not resolve whether the transactions were proper, but it directly contradicts the claim that he furnished no explanation. Majority report, pp. 87–88; transcript, pp. 11, 64, 190–191.

The report also treats charitable entities as possible tools of reputation and payment logistics. Its claim that it found no legitimate activity for the Florida Science Foundation describes the committee’s evidence, not a complete independent audit of every activity or affiliated entity. Entity names, funding sources, grants, loans and operating payments require separate tracing. Tax-planning structures are not intrinsically criminal simply because Epstein used them. Majority report, pp. 91–92.

Deutsche Bank: what the regulator found

The July 2020 NYDFS consent order is a stronger primary basis for banking analysis than an unexplained transaction total. It found that Deutsche Bank failed to tailor due diligence and monitoring to Epstein’s known risks. The relationship ran from August 2013 until the bank decided to terminate it in December 2018 and involved more than forty accounts across Epstein, related entities and associates. The order describes payments to alleged co-conspirators and over 120 wires totaling $2.65 million to Butterfly Trust beneficiaries for stated expenses including tuition, rent and hotels. NYDFS consent order, paragraphs 8, 24, 27–32, 55–57; Majority report, pp. 93–95.

The majority conflates the withdrawal arithmetic. The order describes 97 withdrawals of $7,500 during 2013–2017, totaling $727,500. It separately describes a $100,000 withdrawal in 2018 and a broader cash total above $800,000. The order also states that currency transaction reports were filed appropriately. The documented regulatory failure was inadequate risk-based scrutiny, not that every cash transaction went unreported. NYDFS consent order, paragraphs 48–52; Majority report, p. 94.

NYDFS records both the concerning pattern and the attorney’s denial of structuring. It expressly leaves actual criminal purposes to criminal authorities. Its $150 million penalty covered deficiencies involving Epstein, Danske and FBME, rather than an Epstein-only fine. Suspicious-activity reports and currency reports serve different purposes; filing one does not establish that monitoring was sufficient or that a transaction was criminal. NYDFS consent order, paragraphs 49–57 and penalty provision.

JPMorgan: warnings, relationships and reporting

Staley’s transcript quotes an internal 2006 document describing cash withdrawals of $40,000–$80,000 several times a month, alongside press reporting about the Florida indictment. The majority shortens this to a monthly figure, which changes the scale. Staley acknowledged warning signs while describing a reduced relationship focused on banking rather than managed assets. transcript, p. 122; Majority report, p. 94.

The majority discusses account counts, cross-border movements and retrospective suspicious-activity reporting as signs that warnings were not acted on promptly. Those claims have to be traced to their particular bank records, regulatory findings or pleadings. They should not be merged into a single grand total, and a destination in Russia, Ukraine or another country does not make a transfer criminal. A pleading records a party’s allegation unless a court or other authoritative fact-finder adopts it. Majority report, pp. 94–95.

Universities: donations, access and different kinds of payments

The report argues that philanthropy and academic relationships helped Epstein regain legitimacy after conviction. Its table mixes gifts to institutions, payments for individual students and consulting fees. These are not one uniform category. Institutional reviews supply important corrections and qualifications. Majority report, pp. 96–99.

• Harvard: its review found $9,179,000 in Epstein gifts, with the last gifts in 2007 and none accepted after the 2008 conviction. Continued access was a separate failure: the report documented numerous later visits, office/keycard access and favorable webpage material at the Program for Evolutionary Dynamics. It reported no evidence of engagement with students during those meetings, apart from attendance at one undergraduate class. Harvard review, pp.3–4, 10, 17–19.
• MIT: the university-wide total was ten gifts totaling $850,000 in 2002–2017. Six Media Lab gifts accounted for $525,000. The majority incorrectly assigns the full $850,000 to the Media Lab. MIT’s review found no evidence that Gates or Black donations were actually Epstein’s own funds. MIT/Goodwin report, printed pp.7–11.
• Ohio State: the reported $336,000 includes an unconfirmed $75,000 pledge. The separate $2.5 million COUQ payment was traced by the university to Wexner funds and supported the Woody Hayes Athletic Center, not the arts center. The distinction changes both attribution and purpose. Ohio State’s completed review.
• Bard: the majority’s account needs to be read with the college’s institutional and forensic responses. Bard announced Botstein’s retirement effective June 30, 2026 following an independent review. The majority’s p.96, footnote 555 cites EFTA00989511, one of the two PDFs examined in the Bard-commissioned FTI Consulting report dated October 19, 2026; the other is EFTA01922992 (FTI p.14). The document match is therefore identifiable. FTI said the purported Botstein message appears as quoted text inside an Epstein-to-himself forward and that no native source email with complete delivery, threading and transmission information was available for examination (p.3). It raised substantive concerns about authenticity, integrity and origin but did not establish fabrication with certainty. Botstein denied writing or sending the disputed message. These findings qualify reliance on that exchange without resolving every other record about his relationship with Epstein. Bard, May 1; Bard, October 22; FTI report, pp.3–4 and 14; detailed Wiki source review.

The remaining table lists payments associated with Stockholm School of Economics and Ehnbom, Arizona State’s Origins Project, nursery education, Fashion Institute of Technology, Antilles School, the New School and other education-related activity. Its aggregates depend on the cited bank presentation and underlying records. They cannot all be described as donations, admission inducements or evidence of trafficking. Tuition recipients should not be identified through redacted records merely to complete a table. Majority report, pp. 97–99.

6. International travel, recruitment and intelligence claims

Pages 100–105 connect overseas relationships with practical arrangements: introductions, flights, visas, accommodation and money transfers. The report discusses Andrew, Ehud Barak and other internationally prominent contacts, and describes visa expenses and changes in ledger terminology. The underlying ledgers are cited as committee-held records; the report’s asserted pattern is not a substitute for a transaction-by-transaction public reconstruction. Majority report, pp. 100–105.

The report recounts allegations involving Andrew and Barak. Their denials and procedural context are necessary. Andrew denied the abuse allegation and settled Giuffre’s civil case, which was dismissed with prejudice in March 2022; there was no trial verdict on her claim. The filed settlement statement acknowledged her as an established victim of abuse and expressed regret for Andrew’s association with Epstein, without admitting that Andrew assaulted her. Barak publicly denied having sex with Giuffre. Association, meetings and introductions cannot establish those disputed allegations. Filed Andrew settlement statement; Civil docket; Barak’s reported denial.

The majority interprets “decoy” and refundable itineraries as possible immigration deception facilitated through travel services. Groff gave a different explanation for the particular decoy itinerary put to her: concealment of a visit from the traveler’s parents, not a visa arrangement. That explanation does not settle all other bookings, but it belongs beside the claim. The report also juxtaposes passport-related denials with survivor accounts. Those are conflicting accounts requiring further investigation, not a completed perjury or immigration-fraud adjudication. Majority report, pp. 101–103; transcript, p. 107.

On intelligence affiliations, the majority says witnesses supplied no direct knowledge of Epstein or Maxwell working for an intelligence service. It also records Acosta’s denial of the intelligence remark attributed to him. Those are limited evidentiary findings. Diplomatic connections, a passport, hearsay or a relative’s alleged activities do not prove an intelligence assignment; witness ignorance does not categorically disprove one. Majority report, pp. 103–105; transcript, p. 27; transcript, p. 16.

7. Custody failures and Epstein’s death

What was already established by the inspector general

The custody chapter largely consolidates the June 2023 OIG investigation, supplemented by Noel’s 2026 interview and Barr’s 2025 deposition. OIG documented failures in staffing, supervision, inmate checks, recordkeeping, cellmate assignment and surveillance recording. These left Epstein alone and unmonitored for nearly eight hours, from approximately 10:40 p.m. on August 9 to 6:30 a.m. on August 10, 2019. The report’s minimum of twelve missed rounds and three overnight counts should not be mistaken for a complete tally of every earlier failure. Majority report, pp. 106–114; Majority report, p. 132; OIG, printed pp.94–95, 102–103.

OIG nevertheless found no evidence contradicting the FBI’s determination of no criminality in how Epstein died. The New York City medical examiner determined the manner of death to be suicide. Those conclusions coexist with the documented misconduct. They do not mean no records were falsified, and the existence of institutional failures does not itself establish homicide. Noel and Thomas were charged over falsified records, entered deferred-prosecution agreements and had the charges dismissed after fulfilling their terms; they should not be described as convicted. OIG, printed pp.94–95.

The July incident, cellmate transfer and cameras

The majority calls the July 23 incident an attempted suicide. OIG records conflicting accounts and says the BOP inquiry found insufficient evidence to determine self-harm or assault. Epstein was placed on suicide watch and then psychological observation. The uncertainty in that underlying investigation should not be silently replaced by the majority’s label. The report also compresses arrival and housing dates: SHU assignment was July 7, after his July 6 admission. Majority report, pp. 107–108; OIG, printed pp.29–31 and executive summary.

His cellmate’s August 9 transfer was routine and prearranged, with advance notice documented by OIG. The failure was not an unexplained transfer itself but the failure to replace the cellmate despite the Psychology Department’s requirement. Likewise, a recording-system failure does not mean all cameras had stopped showing live images. OIG describes disk failures that prevented recording for a substantial part of the facility despite continuing live feeds. Majority report, p. 108; OIG, printed pp.51–57; OIG, printed pp.81–85.

Noel’s testimony and disagreements with the earlier record

Noel described understaffing, forced overtime, inadequate training and routine departures from procedure, which she called “the MCC way.” She said she signed a training log without taking the relevant training at a supervisor’s direction, described exhausting shifts, and acknowledged no counts during the midnight shift. These are her recollections. They should be compared with surveillance and other records rather than treated as independent observation by the committee. Majority report, pp. 108–111; transcript, pp. 17–22, 29, 39.

Her account of performing an improper 10 p.m. count is disputed by OIG, which rejected her earlier claim that the count occurred using surveillance, inaccurate slips and other evidence. Noel also denied sleeping in her congressional interview, whereas OIG found apparent sleeping on the footage. The majority’s retelling does not resolve those contradictions. Required rounds and scheduled counts are different procedures and should not be conflated. OIG, printed pp.102–103; transcript, pp. 39, 107.

The orange image and discovery of the body

The report discusses a blurry orange shape near the stairs at approximately 10:39 p.m. Noel could not identify it, denied that it was her and said she did not admit anyone to the SHU. She also could not reconstruct exactly where she had gone off camera or what she could see. These limits prevent her testimony from conclusively identifying or excluding every possibility. The image appears just before the approximately 10:40 p.m. start of the interval used in OIG’s and the FBI’s no-entry account; the image alone does not disprove the official death finding. Majority report, pp. 111–112; transcript, pp. 83–92; July 2025 DOJ/FBI memorandum, p.2.

The majority’s subsection title misidentifies Noel as the officer who discovered Epstein. Noel explicitly said Thomas found him while distributing breakfast. She activated an alarm and went for a cutter; when she returned, Thomas was performing CPR. She said she could not see what he initially saw. Her clarification about a protocol breach concerned opening the cell door, not CPR itself, contrary to the majority’s summary. Her belief that the death was suicide was an inference, not eyewitness testimony of the death. Majority report, pp. 112–114; transcript, pp. 97, 101–102.

8. Disclosure, survivor privacy and the political dispute

What the transparency law required

The Epstein Files Transparency Act required covered unclassified DOJ, FBI and U.S. Attorney records to be released within thirty days in searchable, downloadable form. It prohibited withholding, delay or redaction simply to avoid embarrassment, reputational harm or political sensitivity. Its limited exceptions included victim-identifying/private information, child sexual abuse material, narrowly tailored temporary protection of active investigations, certain images of death or injury, and properly classified information. It also required written redaction justifications and a report to Congress after completion. Public Law 119-38, sections 2–3; Majority report, pp. 119–124.

The majority credits Trump and Congress for disclosure but acknowledges that DOJ did not meet the statutory completion deadline. It contrasts Bondi’s February 2025 “client list” remark and the first-phase binders with the July 2025 DOJ/FBI statement that its review found no incriminating client list, no credible evidence of blackmail of prominent people and no predicate for investigating uncharged third parties. Bondi later said she meant files. These changing official descriptions do not establish that a definitive criminal “client list” existed. Majority report, pp. 117–121; DOJ/FBI memorandum.

DOJ’s January 30 announcement described more than three million additional pages, nearly 3.5 million cumulative pages, more than 2,000 videos and 180,000 images. It said the initial collection exceeded six million potentially responsive pages and attributed nonproduction to duplication, unrelated material, statutory exceptions and privilege claims. Those explanations are DOJ’s account, not an independent reconciliation. The statute’s listed exceptions do not expressly include attorney-client, work-product or deliberative-process privilege, making the claimed legal basis for those withholdings a separate disputed issue. January 30 production letter; DOJ release figures.

Survivor privacy and access to personal records

The report describes deficient redactions, disclosure of sensitive survivor information and difficulty obtaining one’s own records in usable form. Protecting survivors does not require suppressing all accountability material, and public access does not require exposing victims’ identities or intimate information. The report recommends a separate process through which survivors and their representatives can obtain their own unredacted files, rather than indiscriminate unredacted publication. Majority report, pp. 124–125; Majority report, p. 135.

Bondi attributed mistakes to volume and time pressure and relayed an approximately one-percent error estimate. That is testimony, not a verified audit result. OIG announced a compliance audit on April 23, 2026 covering record collection, redaction and withholding guidance, and responses to concerns after release. An audit announcement is not an audit finding of compliance or noncompliance. OIG audit announcement; Majority report, pp. 124–125.

Trump and the disputed political conclusions

The majority asserts that the witnesses and documents it reviewed did not implicate Trump in Epstein’s crimes. It then makes a broader political argument that Democrats focused on Trump rather than survivor justice. These are the majority’s conclusions. Much of the quoted testimony is explicitly limited to a witness’s personal knowledge, recollection or a particular conversation. Bill Clinton’s account of what Trump said, Indyke’s lack of knowledge about a relationship and Wexner’s lack of knowledge do not logically establish universal exoneration. Conversely, an association or an untested allegation is not proof of guilt. Majority report, pp. 125–129.

The majority also disputes how Democrats released photographs, emails and other selections from large document collections. A small selection can lack context, but its size alone does not prove deception. A redacted face does not establish that the person is a minor or survivor. Finding that one photograph shows adults does not establish the identity or circumstances of everyone in other images. The report’s discussion of Plaskett messages and a Jeffries fundraising solicitation concerns contacts and solicitation, not a finding of participation in sexual crimes or proof that a requested donation was made. Majority report, pp. 128–129.

There is a direct disagreement in the congressional record. Oversight Ranking Member Robert Garcia’s November 10 response called the investigation incomplete, criticized the absence of public hearings, and demanded scrutiny of Maxwell’s transfer and Todd Blanche’s role. His assertion that millions of pages remained unlawfully withheld is the minority’s contention, not a completed judicial finding. Separately, House Judiciary Democratic staff issued an October 2026 report arguing that allegations and records warranted further investigation. That is a different committee’s partisan report, not a bipartisan ruling. Neither side’s rhetoric should substitute for the underlying evidence. Garcia’s November 10 response; House Judiciary Democratic staff report.

9. Findings and all six recommendations

The findings on pages 130–132 restate the majority’s central judgments: institutional failures permitted abuse to continue; the 2007 resolution failed survivors; financial and professional relationships helped sustain access; government custody failed; and disclosure remained incomplete and harmful in important respects. Its conclusion calls for further investigation, including follow-up with Groff and investigation of Kellen’s allegations. Those recommendations do not themselves bring charges or determine guilt. Majority report, pp. 130–136.

• Protect survivor and witness speech. Congress should pass legislation such as the Survivor’s Voice Protection Act, H.R.10389, limiting confidentiality clauses that prevent disclosure of sexual abuse and certain supportive disclosures. The introduced bill does not erase every settlement term or all lawful confidentiality. Its retroactive NDA provision and separate fiduciary reporting provisions have different scope and effective-date rules. Introduced H.R.10389, sections 2–4.
• Provide access to personal records. DOJ should create a process for survivors and their representatives to receive their own files unredacted for legal recourse. This is not a call to expose everyone’s private records publicly.
• Interview or reinterview key witnesses. DOJ should revisit people with relevant knowledge, particularly where earlier questioning was absent or incomplete. Whether evidence supports a charge remains a prosecutorial question.
• Apply tighter scrutiny to non-prosecution agreements. DOJ should limit their use to appropriate circumstances, protect survivor interests and require senior review. The report does not propose eliminating every NPA.
• Enforce conflict protections. DOJ and U.S. Attorneys’ Offices should strengthen disclosure, policies and training where needed. A systemic recommendation is not proof that every suspected conflict in the narrative occurred.
• Review financial-reporting compliance. Treasury and FinCEN should examine institutions’ existing reporting obligations and whether regulations need updating, including practices involving prominent clients and delayed reports. A suspicious pattern remains an investigative indicator rather than an automatic criminal finding.

The bill’s proposed fiduciary duty is narrower than a general slogan about “mandatory reporting.” It applies where a fiduciary has reasonable cause to believe the principal engaged in qualifying grooming. The associated civil remedy depends on specified conditions, including subsequent abuse and conviction of the principal. The bill was introduced September 15, 2026; a proposal should not be described as an already-enacted legal requirement. H.R.10389 text; Majority report, pp. 134–136.

American Express travel evidence and regulatory referral

The report discusses American Express records involving decoy and refundable travel arrangements on pages 102, 132, and 136. A related January 14, 2016 email is preserved as DOJ record EFTA02062951, with a searchable copy at epstein-data.com. The document was released on January 30, 2026. It should be read as evidence of a travel arrangement, not as proof that every person named or copied knew of abuse or committed wrongdoing.

The committee recommends that the Treasury Department and FinCEN review financial reporting compliance. The official majority staff report and the committee release document that recommendation. They do not establish that Treasury or FinCEN has opened a new investigation. Tracing the documented operational and financial mechanisms can support accountability without placing responsibility on survivors.

10. Material source corrections and unresolved questions

The differences below affect the substance of the account, not merely copy editing. They show why the report should be read beside its sources.

Other questions remain open in the cited material: the whereabouts and contents of missing computers; the full scope of uncompleted investigative steps; what particular staff and institutions knew at relevant times; the precise provenance and intended meaning of disputed correspondence; reconciliation of bank and estate totals; completeness and legality of DOJ withholding; and whether further interviews will produce new evidence. The majority’s acknowledgment that it had not interviewed everyone with potentially relevant knowledge is important context for any claim of finality. Majority report, p. 7; Majority report, pp. 133–136.

The report does not reproduce every record on which it relies. Some footnotes cite committee-held ledgers or documents “on file,” others link public EFTA records, and others rely on reporting. A quotation in a congressional report verifies what the report says; it does not automatically authenticate the underlying item or settle a disputed interpretation. Page citations here distinguish the majority’s account from directly available transcripts, court decisions, regulatory orders and institutional reviews. This article does not attempt to identify redacted survivors or restore information removed to protect them.

11. Interview and deposition source index

The official release supplies the following records. “Deposition” and “transcribed interview” are preserved as the committee labels them; the existence of a transcript does not mean a witness answered every question. Maxwell’s appearance and Black’s nonappearance should not be counted as full substantive accounts. The date for Staley is inconsistent across the majority schedule and the published transcript cover, so both are identified below. Official transcript collection.

12. Related Wiki coverage and community sources

• U.S. House Oversight Committee
• U.S. Department of Justice
• Sarah Kellen’s congressional interview
• Kellen allegation referral
• Lesley Groff’s congressional testimony
• Richard Kahn’s deposition
• Darren Indyke’s deposition
• Jes Staley’s testimony
• House Judiciary Democrats’ October 2026 report

Community source credit: Epstein Forums administrator Unredacted posted the report and assembled an official transcript-link index on November 10, 2026. These posts are useful discovery and access resources; the transcripts are congressional records, not original transcriptions authored by the forum poster. Report post; Transcript index.

Core supporting records include the OPR executive summary, transcription of the court-filed full OPR report, OIG Report 23-085, NYDFS consent order, Epstein Files Transparency Act and H.R.10389 introduced text. Their different dates, subjects and legal standards are part of the evidence, not interchangeable endorsements of the majority report’s conclusions.