Open the interactive Trump Trades tool on TREMOR - the Annual 2025 view has the tables, the companies rollup and per-table downloads.
About this filing
This section covers President Donald J. Trump's annual public financial disclosure report (OGE Form 278e) for calendar year 2025 - a 927-page filing certified by the U.S. Office of Government Ethics on June 30, 2026. It is a different document from the periodic transaction reports (OGE Form 278-T) that power the rest of this tool.
The annual report is the comprehensive yearly statement: positions held, business entities and their income, the full securities and cash holdings picture, the year's transactions, liabilities, and gifts. The 278-T periodic reports, by contrast, are narrow and timely: they disclose only securities transactions, filed within weeks of the trades.
Important: the annual report does NOT repeat trades already reported on 278-Ts
OGE rules allow a filer to omit, from the annual report's transaction section (Part 7), transactions that were already disclosed on periodic transaction reports. This filing uses that provision: the 21,285 transactions shown here are the 2025 trades that had not previously appeared on a 278-T (largely the systematic buying and selling inside the managed investment accounts). The bond and municipal-bond trades that were disclosed on 278-Ts during and after 2025 are not repeated here.
Practical consequence: neither document alone is the complete 2025 trading record. The full picture is this annual report's Part 7 combined with the periodic 278-T filings (see the Trump Trades methodology). We verified this structure empirically before publishing: the two transaction sets are essentially disjoint. The filing's own cover notes that late filing fees were paid for transactions not previously reported on 278-Ts.
What is in each part
Each part of the filing is a separate table with its own structure - they are not slices of one big table. Full record counts and per-table downloads are on the Data tab of the tool, and on tigzig.com/downloads.
- Transactions (Part 7): 21,285 trades across 8 managed investment accounts - security, type, exact date, amount range.
- Holdings (Part 6): 6,181 asset lines - securities, funds, bonds and cash across the investment accounts, the revocable trust, a family trust and direct holdings - with value and income ranges.
- Business & Income (Part 2, Schedule 1): 501 lines covering the operating businesses, golf and resort properties, licensing and crypto entities. Many lines disclose exact dollar income.
- Liabilities (Part 8): 11 lines and Gifts (Part 9): 11 lines, plus positions (Part 1: 4), employment agreements (Part 3: 3) and spouse assets and income (Part 5: 11).
Ticker matching and confidence ratings
The filing writes security names as free text, with inconsistent spellings and names truncated by the filer's own formatting - the same stock can appear under several variants. We normalize those variants and match each one to a public ticker through a curated mapping table that is shared with the Trade Reports section and grows with each filing. Every match carries a confidence rating, shown as its own column on the Companies tab:
- high - the name identifies exactly one listed security; the overwhelming majority of matches.
- medium - identification is probable but one detail is ambiguous: a share class the filing does not specify, a company renamed or merged during the period, or a truncated name with one clearly-more-likely completion.
- low - a best-effort guess, with the reason recorded; treat with care. Only a handful of names fall here.
- identified, no ticker - the instrument is recognized but is not a listed common equity (preferred stock, notes); shown with a bracketed tag instead of a ticker.
- rule-based - funds, money-market vehicles and bonds classified deterministically by name pattern; these do not need a ticker for the rollup.
Matches were produced in reviewed batches against the normalized name list and reconciled with the shared mapping so both sections of this tool agree on every ticker. Anything not matchable at high confidence is labeled rather than forced - the rating column exists precisely so you can filter or discount the softer matches.
How the data was extracted and validated
The published PDF is a native digital document (not a scan), which allows deterministic, character-exact extraction. Our process, at a high level:
- Two independent extraction methods. The tables were extracted twice, using two unrelated parsing engines and approaches, and the two outputs were reconciled field by field across every row. The final dataset required full agreement.
- Completeness gates. The filing numbers its rows; we verify every numbered sequence is complete - no missing and no duplicated rows anywhere - and that every value belongs to the official OGE vocabulary of amount ranges, transaction types and dates.
- Independent AI vision cross-reads. Samples of pages (and the full business-entity schedule) were independently re-read from page images by multiple AI vision models, blind to our dataset. Every disagreement was adjudicated against a high-resolution image of the printed page before sign-off.
- External tallies. Aggregates computed from the extracted data were reconciled against independently reported figures (major-outlet coverage of the filing), including exact-dollar matches on the largest income items.
No unresolved discrepancies remained at publication. Where the filing itself contains quirks - inconsistent capitalization, occasional typos, security names truncated by the filer's own formatting - we preserve the text exactly as filed rather than "fixing" it.
Accuracy, and what this dataset is
A note on sources, in the interest of full transparency. The filing itself appears to have been prepared as a spreadsheet and printed to PDF - the 927-page certified PDF is what the U.S. Office of Government Ethics publishes, and it is what we parse. OGE also provides copies of disclosure reports on written request (OGE Form 201, under Section 105(b) of the Ethics in Government Act); requesters must identify themselves and statutory restrictions apply to how the reports may be used. No machine-readable version is part of the public posting, so everything here is parsed from the certified PDF.
We have gone to considerable lengths to make that parsing accurate - the independent double extraction, completeness gates, vision cross-reads and external tallies described above, with zero unresolved discrepancies at publication. Even so, this dataset is a best-effort reproduction of a 927-page government document, not the document itself. It is provided as-is for research and general information, without any warranty of accuracy or completeness, and our standard website terms apply. The authoritative record is always the OGE filing itself: the full PDF is linked under Source below, so any figure can be checked against the original. If you believe something here does not match the filing, we want to know - contact details are in the footer, and we will verify against the source and correct promptly.
Reading the numbers
- Securities values and transaction amounts are disclosed as ranges (e.g. $1,001 - $15,000), per the OGE form design. Where we aggregate, we use the range midpoint and say so; open-ended ranges ("Over $50,000,000") use the lower bound. Treat all such totals as approximations.
- Business-entity income in Schedule 1 is often an exact dollar figure - those are shown as filed.
- EIF = Excepted Investment Fund (a pooled fund; underlying holdings need not be itemized and income type is not required).
Source
U.S. Office of Government Ethics, Executive Branch Public Financial Disclosure: official filing PDF (OGE received June 29, 2026; certified June 30, 2026). Extracted July 2, 2026. The next annual report, for calendar year 2026, is due in 2027.
See it live
This page is the static, readable companion to the Annual 2025 view of TREMOR's Trump Trades tool. Open the interactive tool on TREMOR. TREMOR is part of tigzig.com - AI for analytics, databases and macro signals.