Methodology

How each number on the site is produced, and where the data falls short. If something looks wrong, email info@sheepleocrazy.com.

Reports and updates

Members of Congress file a Periodic Transaction Report for covered trades by themselves, a spouse or a dependent child. Sheepleocrazy checks the Senate's eFD site and the House Clerk's site twice a day for new reports, downloads each one, and keeps the original page or PDF. Every trade links back to its report.

The Senate removes reports after about six years, so many from 2012 and 2013 were gone before this site could read them. The House index starts in 2015.

Scanned paper reports

Most reports are filed electronically and read automatically. Some are scanned paper forms. For Senate scans, each trade's type and amount are read from the form's check boxes by image analysis, the names and dates are read with the help of an AI model, and a person reviews every row before it's published. House paper scans aren't transcribed yet, so their trades are missing; their report pages link to the scans.

Amounts

Members report each trade's value as a range, such as $1,001–$15,000 or $50,001–$100,000, not as an exact amount. The site shows the range. Where it adds trades up (as in "estimated volume"), it uses the middle of each range and counts an open-ended range such as "Over $50,000,000" at its lower bound. Treat those totals as rough.

Owners

Each trade shows whose account it was in: the member's own, a spouse's, a joint account or a dependent child's. Trades by a spouse or child count toward the member's totals, as they do in the reports.

Ticker symbols

Reports describe assets in the filer's own words, and many leave out the ticker. Each asset is matched to a symbol in this order: a correction made by hand, the ticker written on the report, a ticker inside the asset name, and finally a close match against the names of listed securities. Matches that aren't close enough wait for a person to check them. Bonds, private funds and other assets without a symbol are listed without one.

Exchanges sometimes give a retired ticker to a different company, so a few older trades may be compared with the wrong price history.

Amendments

A Senate amendment restates the whole report: the amended version replaces the original, and the original's trades stop counting. House amendments usually correct single trades, and the corrected trade replaces the earlier one; a trade the amendment marks as deleted is taken off the site. Replaced reports stay on the site, marked as replaced.

Disclosure lag and late reports

The STOCK Act gives members 30 days from learning of a trade, and no more than 45 days from the trade itself, to report it. Disclosure lag is the number of days from the trade date to the date the trade was first reported, and trades reported more than 45 days after the trade date are marked late. A trade restated or corrected in an amendment counts from the earliest report that listed it; if that report isn't on file, its lag is left blank rather than marked late. A late mark describes timing only: the reports don't always show when a member learned of a trade, so it isn't a finding that a rule was broken.

Timing vs the S&P 500

For purchases and sales of stocks and funds with a known ticker, the site compares the stock's return over the next 30, 90 and 180 days with the S&P 500 (SPY) over the same days, starting from the first daily close on or after the trade date. A purchase scores positive when the stock beats the index; a sale scores positive when the stock trails it, since selling before a relative drop is good timing. The member scores on the Analytics page use the 90-day results and need at least 10 scored trades.

Prices are adjusted daily closes from Yahoo Finance. Options and exchanges aren't scored, and delisted symbols often have no price history. The score ignores position size, because amounts are ranges, and it says nothing about why a trade was made.

Clusters

A cluster is a run of purchases (or of sales) of the same stock by at least three different members, each trade within 7 days of the one before. Popular stocks produce clusters by coincidence all the time, so a cluster is a starting point for a closer look, not a conclusion.

Committee overlap

Committee overlap marks trades in companies whose sector or industry lines up with a committee the member sits on, such as a bank stock traded by a member of a banking committee. The sector-to-committee mapping is a rough rule of thumb. Committee assignments are known from the start of the current Congress (January 2025), so earlier trades are never marked.

AI-assisted summaries

Some member and stock pages have a short summary. An AI model drafts it from the numbers on that page only, and a person checks it before publishing. Every summary is labelled.

Known limits

• Only what members report: trades of $1,000 or less, qualified blind trusts and widely diversified mutual funds generally don't need a report.
• Reports the Senate removed before they were read, and House paper scans, are missing.
• Names, tickers and dates come from reports that sometimes contain mistakes.