What a 13F tells you, in one paragraph
A Form 13F is a quarterly list of the long positions in certain U.S.-traded securities that an institutional investment manager held on the last day of a calendar quarter. Managers whose reportable holdings exceed a threshold set by SEC rule must file it within 45 days after quarter end. It names each security, its CUSIP, the number of shares, and a reported value. It does not show short positions, most derivatives, bonds, cash, non-U.S. listings, or anything bought and sold inside the quarter. Read correctly, a 13F is a dated inventory that tells you what a manager owned on one day, roughly six to thirteen weeks before you read it. Read incorrectly, it becomes a buy list, a timing signal, or a crowd that was never there. The sections below are about staying on the right side of that line.
What Form 13F is and who files it
Form 13F is a report filed with the U.S. Securities and Exchange Commission by institutional investment managers. In general terms, an institutional investment manager is an entity that either invests in securities for its own account or exercises investment discretion over accounts belonging to others. Banks, insurance companies, pension managers, mutual fund complexes, hedge fund advisers, family offices structured as advisers, and broker-dealers with discretionary accounts can all fall inside the definition. What triggers the filing obligation is the size of the manager's reportable holdings: once they exceed a threshold set by SEC rule, the manager files for the quarter in which the threshold was crossed and continues filing according to the rule's schedule. The dollar figure for that threshold is set by rule and can be changed by rule, so this guide does not quote it. Check sec.gov for the current requirement if the exact number matters to your work.
The form has two parts worth knowing. The cover page and summary identify the filer, the reporting period, the type of report, and any other managers whose holdings are included or excluded. The information table is the part most readers care about: a row for each reportable position, with the issuer name, the class of security, the CUSIP, a reported value, the number of shares or principal amount, and a handful of flags that describe discretion and voting authority. Modern filings are submitted in a structured format, which is why data vendors and sites like StockEyes can parse thousands of them each quarter.
Three report types exist. A manager may file a holdings report that lists all of its own reportable positions; a notice that says its positions are reported in another manager's filing; or a combination report that lists some positions and points to another filing for the rest. This matters because a notice filing looks like an empty filing in a naive database query, and a combination report can make one manager appear to hold less than it does unless you follow the cross-reference. When you count how many managers hold a stock, you are really counting how many filings list it, and the report type decides how that count should be read.
The reason the form exists is disclosure, not research. Congress wanted a public view of the concentration of institutional holdings and their effect on the market. That origin explains a lot about the form's shape: it was designed to describe aggregate ownership at a point in time, not to let outsiders reconstruct a manager's trading. Every frustration readers have with the 13F, from the lag to the missing short side, follows from the fact that it was never meant to be a trading record.
What a 13F does and does not contain
The single most useful habit in 13F reading is to say out loud what the form cannot show before you look at what it does show. The table below is the boundary. Everything in the middle column appears in the filing; everything in the right column is invisible to it, no matter how carefully you read.
| Topic | What a 13F contains | What a 13F does not contain |
|---|---|---|
| Direction | Long positions in reportable securities | Short positions of any kind, and therefore the net exposure of any position |
| Instruments | Securities on the SEC's official list of 13F securities, mainly exchange-traded stocks, some ETFs and closed-end funds, and certain options and convertibles tied to listed securities | Most derivatives, swaps, futures, bonds, bank loans, private holdings, and cash |
| Geography | Securities that trade on U.S. exchanges, including U.S. listings of foreign issuers that appear on the official list | Shares listed only on non-U.S. exchanges, even if the same company has a U.S. listing the manager does not own |
| Time | Positions held at the close of the last day of the quarter | Anything bought and sold inside the quarter, the dates of any trade, and the prices paid |
| Values | A reported market value for each position as of quarter end | Cost basis, realized or unrealized gains, or any indication of profit |
| Reasons | Investment discretion and voting authority flags | Why the position exists, who made the decision, or what the manager thinks about the company |
| Clients | The manager's aggregate holdings across the accounts it reports | Which fund or client owns which shares, and whether a position is a client's instruction rather than the manager's idea |
Reading the table: the right column is longer than the middle one for a reason. A manager who holds a stock long and hedges it with an index short, or who owns a convertible bond and shorts the common, or who runs a merger arbitrage book of paired positions, produces a 13F that looks like a plain long bet in every case. The filing is not wrong; it is partial by design.
The official list deserves one more paragraph. The SEC publishes, each quarter, the list of securities that count as 13F securities for that period. If a security is not on the list, a manager does not report it even if it trades in the United States. Newly listed companies, recently converted share classes, and some small issues can be absent from one quarter's list and present on the next, which produces an apparent new position where nothing changed except the list. The current and past lists are at sec.gov/divisions/investment/13flists, and checking a surprising name against the list is a two-minute task that prevents a common error.
The 45-day lag and the holding that is already gone
A 13F covers positions as of the last day of a calendar quarter and is due within 45 days after that day. Most managers file close to the deadline, so the typical reader sees a quarter-end snapshot about six and a half weeks after the fact. If you are reading the filing a month after the deadline, the snapshot is closer to eleven weeks old. Add the time until the next filing arrives, and at the far end of the cycle you can be looking at an inventory from almost four months earlier.
The lag has two separate consequences, and readers often merge them. The first is staleness: the position may still exist, but the price, the company's situation, and the rest of the manager's portfolio have all moved since the snapshot. The second is disappearance: the position may be gone entirely. A manager that bought a stock in the first week of a quarter and sold it in the last week of the next quarter will show the position once, in the filing for the first quarter end, and then show nothing. By the time you see that first filing, the manager may already have sold. There is no way to tell from the filing itself.
The cleanest mental model is to treat each 13F as a photograph with a timestamp. You can compare two photographs of the same manager taken a quarter apart and say, with confidence, that the share count was higher or lower in the second one. You cannot say when during the quarter the change happened, whether it happened in one trade or fifty, whether the manager bought high and sold some back, or what the manager has done since the second photograph was taken. Any sentence that starts with "the manager bought in" followed by a month is an inference, not a reading.
There is a practical corollary for anyone building a research process around these filings. Because the information arrives in waves at mid-February, mid-May, mid-August, and mid-November, the dataset is close to frozen between those waves. Re-reading unchanged filings every week does not produce new evidence. What can change between waves is the set of amendments, discussed below, and your own understanding of the companies involved. Plan your 13F reading around the filing calendar and spend the time between waves on the business research that the filings cannot do for you.
CUSIP versus ticker, and why the mapping can mislead
The information table identifies each security by CUSIP, a nine-character identifier assigned to a specific issue, not by ticker. Tickers are what humans use, so every tool that presents 13F data performs a translation from CUSIP to ticker somewhere in its pipeline. That translation is where a surprising share of 13F errors originate, and understanding it will make you a better reader of any ownership table, including the ones on this site.
The first problem is that one company can have several CUSIPs. A company with two publicly traded share classes has at least two, and a manager who owns both reports two rows. If a tool maps both CUSIPs to the same ticker, the share counts get added together even though the classes may carry different votes, different prices, and different economic terms. If the tool maps only one class, the other disappears from the ownership picture. Neither outcome is wrong in the filing; both are artifacts of translation. When you see a holder count or share total for a company with multiple classes, ask which CUSIPs are behind the number.
The second problem is corporate actions. A stock split changes the share count without changing the position's economic size, so a manager who did nothing will appear to have doubled or tripled its holding across the split quarter unless the tool adjusts for it. A reverse split does the opposite and can look like a mass liquidation. A merger retires one CUSIP and may issue another, so a position in the acquired company vanishes and a position in the acquirer appears, which a careless reader calls a sale and a new purchase. Spin-offs create a CUSIP that had no prior quarter at all, so every holder shows as a new position in the first filing after the distribution.
The third problem is renamed and reorganized issuers. A company that changes its name, redomiciles, or reorganizes under a new holding company often gets a new CUSIP even though shareholders kept exactly what they had. The old CUSIP's last filing quarter and the new CUSIP's first filing quarter look, in a naive comparison, like everybody sold one stock and bought another on the same day. If you ever see a quarter in which a company's entire institutional base appears to turn over at once, suspect an identifier change before you suspect a coordinated trade.
The practical defense is simple. Whenever a change looks dramatic, go back to the filing's own rows and look at the CUSIP and the issuer name as filed. EDGAR search at sec.gov/edgar/search lets you open the original document. If the dramatic change coincides with a split, a merger, a spin-off, or a name change in the company's own filings for that period, the ownership change is mostly or entirely mechanical.
Amendments, restatements, and confidential treatment
A 13F is not final when it is filed. Managers can and do amend. An amendment can be a restatement, which replaces the original filing in full, or an addition, which adds rows that were omitted from the original. The distinction matters when you build a quarter's picture from the data: a restatement should supersede the original, while an addition should be combined with it. Mixing the two conventions produces double-counted positions or missing ones. If a manager's holdings in a stock look impossibly large for a single quarter, check whether an original and a restatement were both counted.
Amendments arrive on their own schedule, sometimes weeks or months after the original deadline. That means the ownership picture for a given quarter is not fixed on the filing deadline; it continues to settle as amendments come in. A table built the week after the deadline and a table built three months later can legitimately disagree. Any site that presents 13F aggregates should tell you the date through which filings were included, and you should read that date as part of the data.
Confidential treatment is a related but separate idea. A manager may ask the SEC to allow it to omit specific positions from the public filing for a period, typically while a position is being built or unwound and public disclosure could harm the manager's clients. Positions under confidential treatment do not appear in the public information table. When the confidential period ends, the manager files an amendment that discloses them. The effect for readers is that the most deliberate, most carefully built positions are sometimes the ones missing from the filing you are reading, and they show up later as what looks like a sudden new position but was in fact an old one. Whether a request was made, and whether it was granted, is not something a reader can infer from the public table. The rules and procedures for these requests are on sec.gov; this guide only notes that the mechanism exists and that it bends the data in a specific direction.
There is one more revision source that is not an amendment at all: the data pipeline. Structured filings are parsed by software, and parsers can misread a value field, mishandle a row with a missing CUSIP, or apply a stale CUSIP-to-ticker map. When a number looks wrong, the order of suspicion should be: identifier mapping first, amendment handling second, parsing third, and only then an unusual trade. Most "surprising" 13F observations resolve at the first or second step.
Reading the fields: shares, value, weight, and change
Each row of the information table carries a reported value and a shares-or-principal figure. The shares figure is the number of shares held at quarter end, or the principal amount for debt-like reportable securities, with a flag that says which one it is. The value figure is the manager's reported market value of the position at quarter end. One caution about units: the unit convention for the value column has not been identical across all years of filings, so when you compare an older filing with a recent one, read the form instructions for each period rather than assuming the columns line up. A tenfold or thousandfold jump that appears across every row at once is a units artifact, not a portfolio event.
From those two fields you can compute the two numbers that matter most, and you should compute them yourself at least once so you know what the tables on any site are doing. The worked example below uses an invented manager, Example Harbor Advisors, and an invented company, Exampleton Industrials, with the made-up ticker XMPL. Every figure is hypothetical and chosen for arithmetic clarity.
Position weight
Suppose Example Harbor's filing for the quarter ended June 30 reports 400,000 shares of Exampleton with a value of 20,000,000 dollars, implying a quarter-end price of 50 dollars. Suppose the sum of the value column across every row in the filing is 800,000,000 dollars. The position weight is 20,000,000 divided by 800,000,000, which is 2.5 percent. That is the share of the manager's reportable long book that this one name represents. It is not the share of the manager's assets under management, because the 13F total excludes everything in the right-hand column of the coverage table above. A manager with a large short book or large non-U.S. holdings has a 13F total that understates its real size, and weights computed against it are correspondingly inflated.
Quarter-over-quarter share change
Now suppose the same manager's filing for the quarter ended March 31 showed 250,000 shares of Exampleton with a value of 10,000,000 dollars, implying a price of 40 dollars. The share change is 400,000 minus 250,000, which is 150,000 shares, an increase of 60 percent on the prior count. Shares are the right field for measuring activity, because the share count changes only when the manager trades (or when a corporate action intervenes, which you check separately). Compare shares to shares, and only then look at value.
Separating price from activity in the value change
The reported value doubled from 10,000,000 to 20,000,000 dollars, but the manager did not double the position. Part of the increase is price. One way to split it: the 250,000 shares already held rose from 40 to 50 dollars, contributing 2,500,000 dollars of the change with no trading at all. The 150,000 new shares, valued at the quarter-end price of 50 dollars, contribute 7,500,000 dollars. Together they account for the full 10,000,000 dollar increase. A table that ranks "biggest increases" by value will put price movers and real accumulators in the same list; a table that ranks by shares will not. Know which one you are looking at.
What you still do not know
Even with all four numbers computed, you do not know what Example Harbor paid for the 150,000 new shares. The quarter-end price of 50 dollars is a valuation convention, not a purchase price. The manager could have bought at 38 dollars in April or at 52 dollars in June, or in several lots across both months, or bought 300,000 and sold 150,000 back. You also do not know whether the manager holds puts against the position, which would be reportable only in certain cases and would appear in a separate row with a put/call flag, or whether it is short a competitor. The arithmetic is exact and the interpretation is still bounded.
The other columns in the row are worth a glance. Investment discretion tells you whether the manager has sole discretion, shares it with another reporting manager, or shares it with someone else; shared positions can be double-reported across two filers, which inflates holder counts. Voting authority is split into sole, shared, and none, and a large position with no voting authority is a hint that the manager is holding for clients who kept the vote. The "other managers" column points to the cross-references discussed under report types. None of these fields changes the arithmetic, but all of them change how much weight the row deserves.
Three classic over-readings
Most of the damage done with 13F data comes from three habits. Each one takes something the filing genuinely shows and stretches it into something the filing cannot show. Naming them makes them easier to catch in your own notes.
A filing lists what a manager owned. It does not say the manager likes those positions today, that the positions were bought rather than inherited through a merger or an index change, that they are unhedged, or that they are sized the way the weight suggests once the invisible parts of the book are included. A long position paired with an undisclosed short is not a bet on the stock. A position held in client accounts under client instruction is not the manager's idea. A position that remains in the filing because the manager is slowly exiting is not an endorsement. The filing is an inventory; a buy list is a judgment, and the judgment has to be yours, made on the company's own filings.
Two consecutive filings tell you the share count on two days roughly thirteen weeks apart. They tell you nothing about the path between those days. Readers routinely write "manager X bought after the earnings drop in May" when all the filing supports is "manager X held more shares on June 30 than on March 31." The purchase could have preceded the drop. It could have been split across the whole quarter. The manager could have bought more than the net change shows and sold some back. If your note needs a date or a price for the manager's trade, the 13F cannot supply it, and you should either find a source that can or write the sentence without the date.
Large financial groups file through several legal entities, and a single adviser can report on behalf of many funds. A stock that shows a dozen holders can, on inspection, have several of them sharing a parent, a research team, or a single portfolio manager. Conversely, one filing can bundle the decisions of dozens of independent fund teams under one name. Holder counts are counts of filings, not counts of independent decisions. When a "cluster" of buyers appears, the first question is how many distinct decision-makers are behind it, and the honest answer is often fewer than the count suggests and sometimes unknowable from the public table alone.
A fourth, smaller trap is worth a line: reading a reported value as assets under management. The 13F total is the value of reportable long positions, nothing more. It is routinely quoted as if it were the size of the firm, which can be off by a wide margin in either direction. If a page on any site, including this one, shows a value total, read it as "reported 13F value" and not as the manager's size.
Read a 13F in ten minutes and end with a note
The goal of reading a filing is not to admire it but to produce a short, dated note that a future version of you can trust. This routine takes about ten minutes once you have done it a few times, and the final step is the only one that creates lasting value.
- Minute one: write the dates down. Record the quarter-end date the filing covers, the date the filing was submitted, and today's date. The gap between the first and the third is the age of everything you are about to read. If an amendment exists, record its date too and note whether it restates or adds.
- Minute two: check the report type and cross-references. Is this a holdings report, a notice, or a combination? Are other managers included or excluded? If the cover page points elsewhere, the information table is incomplete on its own and you should say so in the note.
- Minutes three and four: total the value column and size the book. Sum the values (or read the summary total on the cover page) so you have a denominator for weights. Remind yourself that this denominator excludes shorts, derivatives, bonds, cash, and non-U.S. positions, and write that caveat next to the number.
- Minutes five and six: compute weight and share change for the names you care about. For each position of interest, calculate the weight against the total and the share change against the prior quarter's filing. Use shares for activity and value for weight. If the share change is large, check for splits, mergers, spin-offs, or name changes in the company's own filings for that quarter before you call it a trade.
- Minute seven: check the flags. Look at the put/call column, the discretion column, and the voting authority split for the positions you care about. A row flagged as an option is not shares. A shared-discretion position may be reported by another filer as well.
- Minute eight: look for what is missing. Did a position you expected disappear? Check whether the security left the official list, whether its CUSIP changed, or whether a later amendment adds it back. Absence is the hardest thing to interpret in a 13F and the easiest to over-read.
- Minute nine: list what the filing cannot tell you. Write the specific unknowns for this reading: purchase prices, timing within the quarter, hedges, client instructions, and anything since the snapshot. This list is what stops the note from drifting into a story later.
- Minute ten: write the research note. Three to six sentences. State the facts in the form "as of [quarter end], [manager] reported [shares] of [company], a change of [shares] from the prior quarter, a weight of [percent] of its reported 13F value." State the caveats from minute nine. State one question the reading raises about the company, and name the document you will read next to answer it, which is almost always a filing from the company, not from the manager.
A note written this way is useful a year later because every claim in it is attached to a date and a field. A note that says "smart money is piling in" is not useful a week later, because nobody, including its author, can check it. The discipline is small and it compounds: a folder of dated, bounded notes becomes a record of what the filings actually showed and what you actually thought, which is the only thing that lets you learn from your own reading over time.
How to read the StockEyes ownership tables
StockEyes builds its ownership pages from SEC Form 13F filings. That sentence is the whole caveat: every table on the site inherits every limit described above. The pages show long positions in reportable securities as of a quarter-end date, aggregated across the filings received through a stated cut-off, joined to tickers through a CUSIP mapping. They do not show shorts, hedges, trade dates, or purchase prices, and no table on the site should be read as a buy list, a timing signal, or a count of independent decisions. With that understood, here is what each page is for and how to read its dates.
The per-ticker ownership page
Each covered stock has a page at /stocks/[ticker]/institutional-ownership, reached from the institutional ownership hub. The page states the quarter-end date it describes and the date through which filings were included, and it is organized around that one quarter: the reported totals (number of reporting managers, shares as filed, value as filed), a table of the largest reporting managers with each one's share of the reported total, a comparison of the totals with the prior quarter, a history of the totals across the quarters the site holds, and a list of the CUSIPs that were mapped to the ticker. Read the dates first, then the CUSIP list, then the totals. If the CUSIP list contains more than one identifier, you are looking at a combined figure across classes or across an identifier change, and the quarter-over-quarter comparison should be read with that in mind. The manager rows are filings, so the independence caveat from the over-readings section applies to every count on the page.
The smart-money report
The smart-money report is a quarterly read across all covered stocks for the most recent filing quarter. It groups stocks by what the aggregate filings showed: the largest reported share accumulation and distribution, the widest expansion and narrowing in the number of reporting managers, the largest holders, and a set of flagged moves that are listed as filed but deliberately not ranked. It states the quarter-end date and the filed-through date at the top. Read its rankings as rankings of reported changes between two snapshots, which is exactly what they are. A stock at the top of the accumulation list had a larger reported share increase than other covered stocks between two quarter-end dates; the report does not and cannot say when inside the quarter that happened, at what prices, or whether it was one decision or many.
Fund profiles
The fund profiles page describes the largest reporting managers in the quarter's data using a bounded slice of each manager's filing rather than the whole thing: the reported value in that slice, the number of stocks in it, the top holding and its weight within the slice, and a description of how concentrated or diversified the slice is. The page is explicit that the figures are reported 13F values, not assets under management, and that a filing is a set of positions rather than a return series. Read a profile as a description of what a filing looked like on one date. It is not a ranking of skill, and a large reported value says more about the kind of institution filing than about any decision it made.
Two habits make these pages more useful. First, carry the stated as-of and filed-through dates into any note you write from them, exactly as you would from a raw filing; the site's dates are the filing's dates, not today's. Second, when a number on the site surprises you, go to the source: the screening guide explains how the site's screens are built, and EDGAR at sec.gov holds the original filing, where the CUSIP, the share count, and the value sit in the manager's own row. The site organizes the filings and does the arithmetic consistently. Whether the arithmetic means anything for a given company is the research question, and it stays yours.
Questions about reading 13F filings
What is Form 13F?
Form 13F is a quarterly report filed with the SEC by institutional investment managers whose reportable holdings exceed a threshold set by SEC rule. It lists the manager’s long positions in securities on the SEC’s official list of 13F securities as of the last day of the quarter, with each security’s CUSIP, share count, and reported value. It is due within 45 days after quarter end.
Who has to file a 13F, and what is the threshold?
Institutional investment managers, a term that covers banks, insurers, advisers, pension managers, fund complexes, and similar entities, must file once their reportable holdings exceed a threshold set by SEC rule. The dollar figure is set and can be changed by rule, so check sec.gov for the current requirement rather than relying on a number quoted in an article.
Does a 13F show short positions or options?
It does not show short positions. Most derivatives are outside the form. Certain options on 13F securities are reportable and appear as separate rows with a put or call flag, where the shares figure refers to the underlying rather than to stock owned. A long position in the filing can therefore be hedged or paired with something you cannot see.
How old is the information in a 13F when I read it?
The positions are as of the last day of a calendar quarter, and the filing is due within 45 days after that day. Most filings arrive near the deadline, so the snapshot is typically around six to seven weeks old on the day it becomes public and older every day after. Positions shown may already have been sold.
Why does a stock show a huge ownership change in one quarter?
Check for mechanical causes before assuming a trade: a stock split or reverse split, a merger, a spin-off, a name or domicile change that produced a new CUSIP, a security joining or leaving the SEC’s official 13F list, or an amendment counted alongside the original filing. Most dramatic single-quarter changes turn out to be identifier or data-handling artifacts.
What is an amended 13F, and how should I treat it?
An amendment revises a filed 13F. A restatement replaces the original in full; an addition adds rows that were omitted. Treat a restatement as superseding the original and an addition as combining with it. Amendments can arrive weeks or months after the deadline, which is why any ownership table should state the date through which filings were included.
Can a manager hide positions from its 13F?
A manager can request confidential treatment from the SEC for specific positions, which keeps them out of the public table for a period. When the period ends, the manager files an amendment disclosing them. You cannot tell from the public filing whether a request was made, so a position that appears suddenly in an amendment may have been held for some time.
Can I compute a position weight from a 13F myself?
Yes. Divide the position’s reported value by the sum of the value column across the whole filing. The result is the position’s share of the manager’s reportable long book, which excludes shorts, derivatives, bonds, cash, and non-U.S. holdings, so it is not a share of total assets. For activity, compare share counts between consecutive filings rather than values, since values move with price.
StockEyes provides informational research organization and transparent calculations, not personalized investment advice, valuation conclusions, trade instructions, or predictions. Data described here reflects the sources and dates noted in the article; verify current pricing, features, and data coverage on each provider’s site before subscribing. Consult a qualified professional for advice about your own situation.