Most of Congress's stock-market edge is one household
By Ryan Frigo · October 22, 2026 · Updated nightly; numbers as of October 22, 2026
Everyone knows Congress beats the market. I checked. It does—and most of that edge is one family.
I rebuilt a portfolio from every stock trade members of Congress disclosed since January 2024: 13,465 priced trades by 144 members, including people who've since left office, parsed from House Clerk and Senate eFD filings. Then I put the same dollars into the S&P 500 on the same days. Pooled, Congress returned 63.4%. The same money in the index returned 51.4%. Every number is on the portfolios page →
That's the headline you've seen a hundred times. Here's the part you haven't.
Nancy Pelosi's household—the trades are a spouse's—accounts for $6.9M of Congress's $8.2M edge over the index. That's 85% (rounded from 84.5%). Take that one portfolio out and, at range midpoints, the rest of Congress beats the S&P by 2.3 points instead of 12.0. Take out the top two (Rep. Cleo Fields is second) and it trails the index by 1.9.
One household is 85% of the edge
In dollars, 38 members beat the S&P 500 by a combined $12M and 53 trailed it by $3.8M, for a net edge of $8.2M. Nancy Pelosi's household alone is $6.9M of it.
Nancy Pelosi $6.9MCleo Fields $1.6MMarjorie Taylor Greene $931K35 other members $2.5M53 members who trailed −$3.8M
Leave out the top portfolios and run it again
+12.0 pts All of Congress, pooled: +63.4% against +51.4% for the same dollars in the S&P 500.
pooled return on the members' disclosed trades the same dollars in the S&P 500, same days. The top portfolios are ranked by dollars gained over the index, and each row is a full rerun of the pooled portfolio without them.
How much rides on the dollar ranges
| Trade size read at | Congress vs S&P 500, pts | Largest share of the edge | Without that portfolio, pts | Members ahead |
|---|---|---|---|---|
| Bottom of each range | +13.8 | Nancy Pelosi 75% | +5.1 | 29/67 |
| Midpoint (used above) | +12.0 | Nancy Pelosi 85% | +2.3 | 29/67 |
| Top of each range | +11.6 | Nancy Pelosi 87% | +1.7 | 29/67 |
Every filing gives a dollar range ($1,001–$15,000, $1,000,001–$5,000,000 and so on). Each row reruns the whole model with every trade read at that end of its range; open-ended top ranges use their lower bound in all three. A share above 100% means the other members' gains and losses net out below zero.
Those figures read every trade at the midpoint of its disclosed range, because filings give ranges, not amounts. So I reran everything at the bottom and the top of every range. At the bottom, the household's share is 75% and the rest of Congress beats the index by 5.1 points. At the top it's 87% and 1.7. The exact number moves. The shape doesn't: most of the edge is one family's portfolio. The range table is below, and on the portfolios page →
The trade-by-trade numbers say the same thing. Over the 90 days after a purchase, 43% of the stocks members bought beat the S&P 500 (n=6,530). Of the 67 members with enough trades to rank, 29 beat the index; the median member trailed it by 3.5 points. Every member is ranked here →
So the data doesn't support “Congress trades on secrets.” It supports something less cinematic. A few large portfolios carry the average, and most members are mediocre stock pickers who are also bad at paperwork: 2,836 trades (10%) were reported after the 45-day deadline, and the median trade took 26 days to show up. The 2026 report has the late filers →
Why I built this
Greederboard scores all 539 sitting members of Congress on self-dealing, 0 to 100, from public records. Stock trading is one of six signals, next to net-worth growth while in office (from annual disclosures), campaign money spent on the member's own life (FEC bulk files), money tied to committee work, former staff who now lobby their old office (LDA filings) and missed votes. Every number links to the filing it came from. Party explains 1.2% of the variation in scores. The party test is on the findings page →
528 of 539 members score Low, and nobody scores above 44. That's a finding too.
What I got wrong
AI agents do most of the research and all of the auditing, and they make a lot of mistakes. Before launch I had a separate agent re-check every published row for the 25 highest-scoring members against its sources. The filings held up: 52 of 52 parsed trades and net-worth sums matched the original documents. The research didn't. 38 of 47 researched rows failed.
So I audited every researched row scoring above 15—all 203. 186 failed (92%). Most were sentences a source didn't fully support, OpenSecrets pages the auditor couldn't open, or sector totals nobody could reproduce. 162 rows were rewritten to what their sources support; 24 were withdrawn. Every before/after score is logged at greederboard.com/corrections/.
That 92% is why the audit log is public. It came from a sweep aimed at the riskiest rows, so it overstates the typical error. The fairer number is the random daily sample: 28 of the 66 rows it has checked failed (42%). Still bad. Each failure gets fixed or withdrawn, the failure patterns feed a weekly methodology review, and every rule change ships with a version number. I'd rather show you the error rate than ask you to trust me.
The rules
A score is an opinion computed by a published formula. It doesn't say anyone broke a law. Missing data is never treated as wrongdoing. Members and their offices can reply on their own page, and statements are published in full. The site takes no money from anyone whose score could change and won't ever sell trading signals. The formula and the sources →
Look up your representative. Then tell me what's wrong.
These are returns on disclosed trades, priced from public market data, not a finding that any member broke a law or traded on nonpublic information. Past returns say nothing about future results.