House disclosure gaps: Q3 2026

The 161 rows above the comparison threshold occur in 7 of the 72 measured filings (9.7%). Three filings account for 151 of those 161 rows (93.8%). Each row has equal weight, so a report with many transactions affects the row percentage more than a report with one transaction.

All figures use the filing period, calculation date, calendar-day method and 45-day comparison stated above. The full cohort remains in the denominator, including same-day filings and the longest gaps.

An amendment's filing date can be much later than the first disclosure of the transaction. Treating it as a new trade can therefore overstate the time before the transaction was first reported.

We kept 8 purchase or sale rows marked Amended separate. Their median transaction-to-amendment gap was 521 days, with a full range of 7 to 547 days; 7 of 8 exceeded the same 45-day comparison. We also excluded 1 row marked Deleted. These labels come from the original forms, not an inference from a ticker and date. The figures describe amendment timing only, over the same Q3 filing period and October 1 calculation date. They do not measure the first disclosure date.

For example, filing 20035035 labels its first transaction Amended while later rows are New. The distinction is printed directly under each asset.

The House Ethics Committee's guidance requires covered transactions to be reported by the earlier of 30 days after awareness or 45 days after the transaction. A simple transaction-to-filing gap does not establish an enforcement finding. Source errors, prior disclosures and amendments need individual review. A gap of 45 days or less also does not prove that the awareness-based deadline was met.

The Clerk's index gives a filing date. It does not establish the exact time that a document first became publicly accessible. We therefore call this a filing gap, not a measured public posting delay.

Filings can describe a member's, spouse's or jointly owned account. A row is not proof that the member personally selected the investment. “New” is the label on the form; it does not independently prove that no earlier disclosure exists.

This is a Q3 filing-date study. It includes transactions from earlier periods that were reported in Q3. It is not a complete study of trades executed during Q3: September transactions can appear in later filings.

Start with the Clerk's official disclosure portal and 2026 filing index.

Select index entries with filing type P and filing dates from July 1 through September 30, 2026, inclusive.

Extract the reported transaction date, row-status label, transaction code and asset-type code from each text-readable original PDF.

Include purchase and sale rows with asset types ST, ET, RS or OP. ST includes some preferred securities. Preserve separate rows, even when their ticker and date match. Do not infer unique trades from identical descriptions.

Measure New rows separately from Amended and Deleted rows. Include large managed-account filings; the product's bulk-filing filter is not a reporting-deadline exemption.

Subtract calendar dates. Do not truncate a stored timestamp into UTC and assume that it is the source's filing date.

Download the row CSV and coverage manifest above. The manifest lists every attempted document and its content checksum, including the unmeasured documents.

The second extraction method matched all 892 rows returned by the existing application PDF parser and recovered 5 additional preferred-security rows. The independent extraction also preserved a Deleted label that the application's purchase/sale output does not represent. No production data was changed.

Limits: Senate disclosures, unmeasured PDFs, unfiled transactions and later additions to the Clerk's index are outside this result. Other asset types and exchanges are outside the headline cohort. This is an observed sample, not a claim of complete House coverage. No stock prices, portfolio returns, insider-trading conclusions or investment recommendations were calculated.

Research contact: vansh@alphayou.net. Corrections should identify the filing and row so the source can be checked and the revision recorded.