Did Copying Congress Beat the Market? What the Data Shows
This is the question behind most interest in these filings: if you bought whatever Congress bought, would you have done better than the market? Here is what the data on this site shows, and how to read it.
The headline numbers
Across 4,700 scored Congress buys, the average return over the 90 days after entry was +0.4% against the S&P 500, with 50% of buys ahead of the index and a t statistic of 0.9. For insiders the average was +1.2% across 1,900 scored buys, with 53% ahead.
These figures are rebuilt from the data every time the site is built, so they move as new trades mature. They were current as of 2026-09-23 (23 September 2026) for Congress and 2026-09-28 (28 September 2026) for insiders.
How to read the t statistic
The t statistic compares the average gap with how much individual results vary. If it is smaller than about 1.65 in size, a gap this large shows up by chance quite often, and you cannot tell it from luck. Between about 1.65 and 2.58 the result is suggestive but could still be chance. Above 2.58 it is unlikely to be chance.
Notice that a result can be positive and still not meaningful. An average of a fraction of a percent on thousands of trades with wide variation is a small effect that may not survive costs.
What “start the day after the filing” does
Every return here starts on the first trading day after the filing, not on the trade date. That is the earliest an outsider could act. Starting on the trade date would show much better numbers, because it counts the run up before the public knew, and no follower could have captured it.
What is not in the numbers
- Costs. Commissions, spreads and taxes reduce real returns.
- Risk. The comparison is with the S&P 500 only.
- Delisted stocks. Stocks with no price history are not scored, which can flatter results.
- Skill versus luck for a single person. The averages above pool everyone. Individual records are far noisier, which is why the leaderboards are ranked out of sample.
The takeaway
Read the sample size before the average, and the t statistic before either. If the data does not show a clear edge, that is a result, not a failure. The full tables, by horizon and by year, are on the performance page.
This is information, not investment advice.