Congressional Trades as an Alternative Data Signal

Members of the US Congress trade stocks, and under the STOCK Act they must disclose each transaction publicly. Those disclosures are a genuine alternative-data source: public, structured enough to parse, and tied to people whose decisions can intersect with the companies they trade. Prismetric ingests the filings and turns them into a signal. This article explains what the data is, how we read it, and — just as important — where it falls short. It is educational, not investment advice.

What the disclosures contain

Each Periodic Transaction Report (PTR) lists a filer, a ticker, a transaction type (buy or sell), an amount band, and the transaction and disclosure dates. The reports are filed as documents, not a tidy database, so the first job is extraction: pull the trades out, normalize tickers, and map each to a company Prismetric already tracks. Trades in names we don't cover are stored but don't drive a signal; the ones that map become part of that entity's picture.

From filings to a signal

Individual trades are noisy — one lawmaker buying one stock means little. The signal lives in the aggregate. Prismetric sums disclosed buying and selling per company over a trailing window to get a net disclosed flow, then folds it into the momentum component of the score, where it is:

• Point-in-time — stamped to when the information became public, so a backtest never sees a trade before it was disclosed.
• Recency-decayed — a disclosure from last week weighs more than one from months ago.
• Rank-normalized — scored as a percentile against peers, so the raw dollar amounts (which are only reported in broad bands anyway) don't distort the cross-section.

On a company's page this also surfaces as a "who's holding this" view, and across the platform the politicians page ranks the most active and most-followed filers so you can see the behavior behind the number.

The limitations are the headline

Congressional-trade data is popular, heavily hyped, and routinely misunderstood. Be honest about what it can't do:

• The disclosure lag is large. Filers have up to 30–45 days to report. By the time a trade is public, the move that may have motivated it can be long over. This is disclosed information, not a live tip.
• Amounts are banded. Reports give ranges, not exact sizes, so any dollar figure is an estimate built on wide bands.
• It is lumpy and sparse. Most companies have no congressional trades in any given window; the signal is informative for a minority of names and silent for the rest.
• Correlation is not a motive. A lawmaker's trade may reflect a financial advisor, a blind-trust-like arrangement, or ordinary rebalancing — not any special knowledge.

How to use it well

Read congressional-trade flow as one corroborating input among many, not a standalone trigger. Its best use is as context: when disclosed buying lines up with other signals on a name you already follow, it adds a small amount of confidence; when it is the only thing pointing somewhere, discount it heavily given the lag. For the broader question of how to combine sparse, delayed sources into something tradable, see measuring adoption with alternative data. Nothing here is personalized financial advice.