Gilbert Cisneros

How the committee overlap is counted, and what it is not

A trade counts toward the overlap when the company's SEC industry classification falls plainly inside what one of the member's committees covers — a bank trade for someone on Financial Services, a defence contractor for someone on Armed Services. The mapping between committee and industry is curated and deliberately short, covering only pairs nobody would argue about. It is not official; no official version exists.

So the number is an overlap between two public records: the House Clerk's disclosures and the committee rosters published by the unitedstates project. It is not a finding, and it carries no claim that the seat and the trade are connected. Two things it will not tell you apart: a member trading a sector because they sit on its committee, and a member sitting on a committee because they spent a career in that sector. The second is extremely common and is how committee assignments are supposed to work.

It also cannot see who actually made the trade. Disclosures cover a spouse's and dependent children's accounts, and many are made by a manager under an arrangement the member does not direct.

What the disclosure record does and does not contain

• Amounts are ranges. The STOCK Act requires a band, not a figure, so there is no portfolio value here and no total. Any site showing you a member's exact position size computed it by assuming something.
• It lags. A transaction must be reported within 45 days of notice, and the notice itself can follow the trade. A date here is the transaction date as filed.
• Tickers are free text. Filers type them by hand into a PDF, so some are wrong or missing and some entries are not securities at all. Only values that look like real symbols are linked.
• Late and amended filings happen, and the record keeps both.

Where this comes from

Disclosures are parsed from the Clerk of the House's own published filings — the full record with every original PDF is on the congress trading pages. Committee rosters and member identity come from the unitedstates/congress-legislators datasets, which are public domain. Neither is bought from a vendor.