How to Read a 13F Filing (and What to Look For)

Four times a year, every institutional investment manager overseeing at least $100 million in US-listed securities must tell the SEC what it owns. The disclosure arrives as Form 13F, due within 45 days of each quarter's end, and it is the closest thing retail investors get to looking over the shoulder of Warren Buffett, Seth Klarman, or Bill Ackman. The trouble is that a raw 13F is just a table: security names, share counts, and market values, with no commentary and no context. Knowing which numbers matter, and which ones quietly mislead, is the whole game.

Start with the holdings table

Whether you read the filing on the SEC's EDGAR system or through a tracker like this one, the core object is the same: a list of positions as of the last day of the quarter. For each position you will typically see four things.

• Market value. The dollar value of the position at quarter-end prices. Impressive on its own, but almost useless without context, for reasons covered below.
• Percent of portfolio. The position's market value divided by the total value of all reported holdings. This is the number professionals look at first.
• Share count. How many shares the manager held. On its own it tells you little, but compared against the prior quarter it reveals whether the manager was buying or selling.
• Change versus last quarter. Most trackers compute this for you: shares added, shares sold, or a position that appeared or vanished entirely.

Two things a 13F will never show you: short positions and most non-US listings. A fund that looks wildly bullish on paper may be hedged in ways the form does not capture. Keep that limitation in mind before treating any filing as a complete picture of a manager's book.

Percent of portfolio beats dollar value

A $500 million stake sounds enormous. For a fund managing $80 billion, it is a rounding error, roughly 0.6% of assets, and it may have been initiated by a junior analyst as a starter position. Meanwhile a $200 million stake inside a $1.5 billion fund is over 13% of everything the manager owns. That is a statement.

Portfolio weight is the honest measure of conviction because it captures opportunity cost. Every percentage point allocated to one stock is a point not allocated to the manager's next best idea. When you scan a filing, sort by percent of portfolio, not by dollar value, and pay special attention to anything above roughly 5% of the book. You can see this clearly in Berkshire Hathaway's portfolio, where a handful of names have dominated the reported equity book for years while dozens of smaller lines come and go with little consequence.

The four types of quarterly change

Comparing one quarter's filing against the previous one turns a static snapshot into a story. Every position falls into one of four buckets, and they are not equally informative.

Change type	What happened	How informative it is
New position	The stock appears for the first time	High. A deliberate decision to deploy fresh capital.
Added	Share count increased	High when the addition is large relative to the existing stake, especially if the stock fell during the quarter.
Trimmed	Share count decreased but the position remains	Low to moderate. Often just rebalancing after a run-up, tax management, or funding redemptions.
Sold out	The position disappeared entirely	High. A complete exit is a clear change of mind.

New positions and full exits are decisions. Trims are frequently housekeeping: a stock that doubled will mechanically grow as a share of the portfolio, and cutting it back says nothing about the manager's view of the business. One useful refinement: an addition made while the stock was falling suggests genuine conviction, since the manager chose to average down rather than walk away. You can browse these changes across every tracked fund on the activity feed, which is far faster than diffing filings by hand.

Concentration is a conviction signal

Two funds can hold the same stock at the same weight and mean very different things by it. A manager running 12 positions has done deep work on each one and cannot afford to be casually wrong. A manager running 400 positions is expressing a portfolio-construction view, or simply hugging an index, and any single name is close to noise.

As a rough guide, a portfolio where the top ten holdings account for well over half of reported assets reflects a concentrated, high-conviction style. When a manager like that initiates a new top-five position, it is worth understanding why. When a 600-name quant fund adds the same stock, it may reflect nothing more than a factor model rebalancing. Same line item, entirely different meaning.

Cluster buying: when unrelated managers agree

Any single filing can mislead you. A pattern across many filings is harder to dismiss. When several managers with different styles, different home bases, and no shared parent company all initiate or add to the same stock in the same quarter, each acted on independent research and reached the same conclusion. Academic work on 13F data has generally found that this kind of overlap among skilled managers carries more signal than any individual holding.

The caveat is independence. Five Tiger-descended funds buying the same software name is closer to one decision than five, since those managers share training, networks, and often ideas. Look for agreement across genuinely unrelated shops: a deep-value contrarian, a quality-growth compounder, and an event-driven fund converging on one ticker is interesting. You can screen for exactly this pattern on the consensus buys page, or take a broader view with most bought stocks for the quarter.

The 45-day lag, and why 13F prices are not entry prices

This is the mistake that costs beginners real money. A 13F covering the quarter that ended March 31 can legally be filed as late as mid-May, and most large funds file at or near the deadline. By the time you read it, the positions are at minimum six weeks stale, and the trades themselves could have been executed at any point during the quarter, up to four and a half months before you see them.

Three consequences follow. First, the manager's cost basis is unknowable from the filing; the market value shown is a quarter-end mark, not a purchase price. Second, the position may already be partially or fully sold. Fast-trading funds turn over ideas in weeks, so their filings describe a portfolio that no longer exists. Third, the stock may have moved sharply since quarter-end, which changes the risk-reward even if the thesis is intact. Treat a 13F as a starting point for research, never as a signal that a price is endorsed. If a stock is up 40% since the quarter closed, you are not buying what the manager bought.

Match the manager's horizon to yours

The reporting lag is fatal for some filings and nearly irrelevant for others. A long-horizon value investor who holds positions for five to ten years is barely affected by a 45-day delay; the filing you read in May still describes the portfolio in August. Filings from managers with low turnover age well, which is why the classic buy-and-hold investors remain the most useful ones to follow.

By contrast, following a high-turnover fund through 13Fs is like navigating with a photograph of where a car was parked last month. Before anchoring on any filing, check the manager's typical holding period, which you can infer from how much of the portfolio changes each quarter. Browsing the manager list with an eye for low-turnover, concentrated funds will save you from chasing ghosts.

Red flags and false signals

A few recurring traps are worth committing to memory.

• Amendments dressed up as news. Managers sometimes file a 13F-HR/A, an amendment that restates or adds to an earlier filing. A poorly built tracker can display a restated holding as a brand-new buy. If a "new" position seems out of character, check whether the source document is an original filing or an amendment.
• Tiny tracker positions. Managers sometimes hold a token stake, well under 1% of the portfolio, simply to keep a company on their radar or to receive shareholder materials. These are watchlist entries, not bets. Filter them out before drawing conclusions.
• Index-like portfolios. A filing with 500 or 1,000 names and weights that resemble a benchmark tells you about index construction, not stock selection. The individual holdings of such funds carry almost no information.
• Quarter-end window dressing. Because the form captures a single day, a manager can buy a fashionable winner just before quarter-end to appear on the right side of it. Positions that appear at small weight and vanish the following quarter deserve skepticism.

None of this makes 13F data useless. It makes it a research tool with known distortions, which is true of every data source in investing. Used carefully, filings tell you where experienced, well-resourced investors are putting real capital, and cross-referencing their picks against your own work, for instance through the stock screener, is a legitimately good way to build a watchlist. Just remember what the form is: a delayed, long-only snapshot of one day per quarter. Read it for the decisions it reveals, not the prices it shows, and let the manager's conviction, consistency, and time horizon carry more weight than any single headline position.