If you’ve ever wondered how people find out what Warren Buffett or Michael Burry is buying, the answer is 13F filings. It’s a mandatory SEC disclosure that gives the public a window into institutional portfolios. Not a perfect window, but a window.
Who files them
Any institutional investment manager with at least $100 million in qualifying assets under management. That includes hedge funds, mutual funds, pension funds, insurance companies, and bank trust departments.
We’re talking about thousands of filers. Not just the big names you see in the news. We track 11 of the most notable funds, but plenty of smaller, under-the-radar firms file 13F reports that almost nobody reads.
What’s in a filing
A 13F lists the manager’s long positions in U.S. equity securities at the end of each quarter. That means stocks and certain convertible securities.
What it does not include: short positions, cash, bonds, foreign stocks, private investments, or options (except for their underlying equity holdings). So you’re seeing a partial picture. A meaningful partial picture, but partial nonetheless.
Each position entry includes the issuer name, CUSIP (security identifier), share count, and market value. That’s it. No explanation of why they bought it, no target price, no time horizon. Just the raw positions.
The timing problem
13F filings are due 45 days after the end of each calendar quarter.
| Quarter ends | Filing deadline |
|---|---|
| March 31 | May 15 |
| June 30 | August 14 |
| September 30 | November 14 |
| December 31 | February 14 |
That means a trade made on January 2nd might not be visible until May 15th. Over four months of delay in the worst case.
And here’s the nuance: the filing shows positions as of the quarter-end date. A manager could have bought in January, sold in February, and the filing would still show the position as of March 31st. You wouldn’t know about the February sale until the next 13F comes out.
This is why simulated copy-trade returns are always lower than the manager’s actual returns. You’re working with stale information by design.
How we process them
We pull 13F filings directly from SEC EDGAR. The filings come in as XML or text documents. We parse the position data, match it against our ticker database, calculate portfolio weights, and compare quarter-over-quarter to identify new positions, exits, and size changes.
Across the 11 funds we track, that’s 232 active positions and 3,619+ processed trades. From concentrated 2-position portfolios (Mohnish Pabrai: 82.7% RIG, 17.3% VAL) to diversified 15-position ones (Buffett: KO 15.9%, AXP 14.0%, CVX 13.3%…).
Using 13F data in practice
Here’s what the actual performance looks like across our tracked funds, all with the 45-day delay built into the backtest:
| Manager | 1Y CAGR | Sharpe | Concentration |
|---|---|---|---|
| Mohnish Pabrai | +116.33% | 1.19 | 2 positions |
| Michael Burry | +28.50% | 0.79 | 4 positions |
| Carl Icahn | +17.04% | 0.44 | 11 positions |
| Pat Dorsey | +15.90% | 1.53 | 7 positions |
| Bryan Lawrence | +14.02% | 1.41 | 9 positions |
| Chris Hohn | +9.13% | 1.05 | 8 positions |
| Buffett | +7.14% | 0.94 | 15+ positions |
| Ackman | +3.09% | 0.63 | 9 positions |
The pattern: more concentrated portfolios tend to have higher returns but also higher risk. Pabrai’s +116% comes from a 2-stock portfolio — incredible when it works, devastating when it doesn’t. Dorsey’s +15.9% with a Sharpe of 1.53 is arguably the best risk-adjusted performance.
Warren Buffett’s portfolio is a good example of long-term conviction positions where the delay matters less. Bill Ackman’s concentrated bets are a different story — the delay can mean you’re buying at a significantly different price.
For how the 13F disclosure system compares to what politicians are required to file, our congressional trading guide covers both systems side by side.
The bottom line
13F filings are one of the best free data sources in finance. They let you see what the smartest institutional money is doing. But they come with real limitations: delayed data, incomplete pictures, and no context around the “why.” Understanding those limitations is what separates useful analysis from blind copying.
